For Jacob — a college student building a construction management app with Claude Code — it was this:

“You'll know you're actually getting good at this when you spend 80% of your time writing the spec, not the code.”

He'd been spending most of his energy wrestling with the code. Debugging, figuring out why Claude produced the wrong output, and iterating. Normal when you're starting out. But I was describing where he'd be in a few months.

The ratio flips.

Why the 80/20 Inverts

Beginners treat AI coding like a shortcut for code generation. Feed in a vague idea, hope something usable comes out, fix what's broken.

That's why beginners spend 80% of their time on code. Not because coding is the important part, but because the vague input created a vague output, and now you're paying the cleanup tax.

Experts do the thinking upfront. They spend most of their time writing what I'd call a PRD — a Product Requirements Document, but that sounds corporate. It's really just a clear, complete description of what you're trying to build.

What should happen when the user clicks this button? What's the edge case when someone enters a letter instead of a number? What does “success” actually look like for this feature?

Answer all of those first. Then hand it to Claude Code.

A good spec is most of the work. The code is almost just output.

Garbage In, Garbage Out — But Not How You Think

I've been saying this for a while: garbage in, garbage out applies to AI more than anywhere else.

People usually take that to mean “use better prompts.” That's true, but it's the wrong level.

The problem usually isn't word choice. It's fuzzy thinking before you start.

I had a coaching session with someone building a construction tracker app. He kept getting messy output — wrong calculations, confused logic. We looked at his prompts. They seemed fine. So I asked him to walk me through the math out loud.

He stumbled.

That was it. He didn't fully understand the problem himself. And if he couldn't explain it clearly to me, he couldn't explain it clearly to Claude.

There's a quote I come back to a lot: defining the problem is half the battle — and it's harder than most people think.

The quality constraint isn't the AI. It's the clarity of what you're feeding it.

What Goes In a Good Spec

This doesn't have to be a twenty-page document. For most small builds, a good spec is just a few paragraphs that answer the following:

What is this for? One sentence: who uses it, and what problem does it solve.

What should happen in the normal case? Walk through the main flow like you're explaining it to someone who's never seen your app.

What are the edge cases? What happens when something goes wrong, or when input is unexpected?

What does done look like? What can you test to confirm it's working?

That's it. You don't need a formal template. You just need to think it through before Claude starts building.

When I flip the script and let AI interview me — ask “what do you need to know to succeed?” — it forces me to answer the questions I'd otherwise skip over. And when I answer them, the output is exactly what I needed.

The Architect Analogy

An architect who's great at their job doesn't build the house.

They draw incredibly detailed plans. They think through every room, every load-bearing wall, every pipe. They anticipate problems before construction starts. That's where their expertise lives.

The builders execute from those plans. They're skilled. But the thinking was already done.

AI coding works the same way. Once you have a clear spec, Claude Code is the builder. It's skilled. It can execute. But it can only work from the plans you give it.

Your job — as you get more advanced — shifts from “making the code work” to “making the plans clear.” That's the real skill development. That's the 80/20 flip.

This Generalizes Beyond Coding

The principle isn't just about code.

Garbage in, garbage out applies every time you ask an AI to produce something — a marketing email, a research summary, a draft proposal. The output quality is downstream of input quality.

Clear context beats clever wording, every time. The better you get at managing what you give the AI — and the more time you spend on that clarity before you start — the better everything downstream becomes.

It's a skill anyone can develop. And it starts with slowing down before you start, not speeding up once you're stuck.

Want to build this skill hands-on? The 4-Day AI Sprint is the fastest way to go from vague prompts to results that actually work.

Eight months.

You know what he had to show for it? Barely anything. The basics of the basics. No project. No working thing he'd built. Just a mental model with no real foundation under it.

Six weeks into our sessions together, I looked at what he'd built — a real construction management app, branching workflows, Docker setup, pull requests — and told him: “That's already intermediate-level development.”

He thought I was being generous. I wasn't.

Same person. Same brain. Completely different result. The only thing that changed was the method.

The YouTube Trap

Here's the thing about tutorial culture: it feels productive.

You're taking notes. Watching a pro explain how things work. You finish an episode and feel like you learned something. That feeling is real. But it's not the same as actual skill.

Real skill — in coding, in AI, in anything technical — only shows up when you're trying to build something specific, hitting a wall, and figuring your way out. That's the part YouTube can't simulate.

I ran into the same pattern when I was experimenting with building Python agents. I'd read articles, watch demos, and still feel vague about what I was doing. The moment I sat down and actually built one — even a basic one that cost $40 a month vs. the $200 I was paying for a no-code tool — everything clicked. The doing taught me things the watching never could.

What Hands-On Learning Actually Looks Like

In Jacob's case, each session started with a goal. Something small and concrete.

First week: set up a GitHub repo, push a real commit.
A few weeks in: add branching and learn what a pull request actually is.
Later: connect Docker, run the app locally, write basic tests.

None of it was glamorous. A lot of it involved breaking things and fixing them. But that's exactly what builds the skill.

With Claude Code as his coding partner, Jacob didn't need to know everything upfront. He'd describe what he wanted. The AI would generate a draft. Then he'd have to read it, test it, understand it well enough to know when it was wrong, and iterate from there.

That feedback loop is the whole thing. That's where the learning happens.

The AI Fluency Levels

When I teach workshops, I explain AI fluency in three stages.

The first stage is AI Assisted — you're using chat tools like ChatGPT for individual tasks. Writing help, research, answering questions.

The second stage is AI Workflows — you're chaining tools and building repeatable processes. Claude Code falls here for most people. You're not just asking questions; you're building something.

The third stage is Building Agents — fully automated workflows running on their own, with minimal human input.

Most people trying to “learn AI” stay stuck between Stage 1 and Stage 1.5. They're using chat tools occasionally, but they haven't crossed into actually building workflows because they're still consuming content instead of making things.

Jacob skipped straight to Stage 2 in six weeks. Not because he's unusually talented. Because he was actually doing.

The Passive Consumption Trap (It's Not Just YouTube)

Last year I imported 300 Lex Fridman podcast episodes into NotebookLM using a Chrome plugin. In minutes I had access to six hundred hours of content. I could ask anything — what tools does he mention most, what are his core ideas, what's his philosophy.

That's useful. But useful in a specific way: it speeds up passive research. It doesn't replace the experience of actually trying something, shipping it, and learning from what breaks.

The mistake most people make is treating AI tools as yet another passive media format. Another podcast to listen to. Another newsletter to read. Another demo video to bookmark.

The acceleration comes from using AI to build, not just consume. Every session Jacob and I had together was him building. The AI was the coding partner. I was there to keep him from getting stuck for too long and point him toward what to work on next.

How to Get Out of YouTube Mode

If you're in the passive learning loop right now, here's how to break out.

Pick one thing you actually need. Not a tutorial project. Not a demo. Something you'd use in your real life or your business — even if it's small.

Then try to build it, even badly. Ask Claude Code for help. Get something working. Then make it better.

The first version will be ugly. That's fine. Ugly and working beats polished and imaginary every time.

Jacob's first version of his construction app was rough. But it ran. And that was the moment things started compounding. Each week he built on the previous week. After six weeks, he had something real.

The YouTube version of that story would have taken years. Or never.

If you want a faster path to AI fluency, check out the 4-Day AI Sprint. It's the hands-on version — building real workflows, not watching demos. That's the difference.

Not wrong in intent. Signs at checkout. Scripts for the receptionist. Follow-up emails. The problem is the timing and the person.

I picked this up from Chris Murphy, who has spent years studying what actually drives review volume in the salon industry. His finding: the most important person for driving 5-star reviews is not the front desk. Not the owner. It's the stylist.

Why the Moment Matters

Think about what actually happens during a salon visit. A client sits in the chair, spends an hour or two with their stylist. Then the stylist finishes, holds up the mirror, and the client sees the result.

That's the peak. Maximum satisfaction. Feeling freshest and strongest.

Then they walk to the front desk. Pay. The feeling is already starting to fade. Two days later, when the review request email arrives, the feeling of that appointment is mostly a memory. That's why conversion is so low.

The Right Ask, at the Right Moment

What Chris found is that the stylist asking right after the reveal converts dramatically better than anything that comes afterward. The stylist, in the moment: I'm so glad you love it. Would you mind leaving us a quick review?

The emotional high is fully present. The personal relationship is right there. The friction to say yes is almost zero.

This plays out across service businesses. The personal trainer after you hit a PR. The aesthetician when the client first sees their skin after a facial. The contractor at the final walkthrough. The person doing the actual work is present at peak satisfaction. That's when the ask works.

The Problem With Making Stylists Remember

The obvious move is to train your stylists to ask. This works, but it's inconsistent. Some feel awkward. Some remember with some clients and forget with others. Over time the training fades.

This is where automation becomes valuable. A trigger fires 20-30 minutes after an appointment ends, while the feeling is still strong. The message comes through as a text with the stylist's name on it, personalized to feel like it came from that person.

Something like: It was so great to see you today! I loved how your color came out. If you have a moment, a quick Google review would mean so much. Not a generic thank-you message. A message that feels personal and references the actual service.

Building the System

Your booking system knows when appointments end. That's the trigger. An AI-drafting tool pulls in the client name, stylist name, and service type and generates a personal-feeling message. The stylist doesn't write anything. The system fires for every appointment automatically.

The 80-20 principle applies: review requests are high-frequency and high-impact. Reviews compound over time and directly affect search visibility and new client acquisition. A salon doing 100 appointments per week converting even 10 percent to reviews adds 10 reviews per week. That's a meaningfully different Google Maps profile within months.

The Broader Principle

The insight from Chris Murphy is about identifying where the emotional peak is in your service and making sure your ask lives there.

Most businesses design their review strategy around what's convenient for the business rather than what's optimal for conversion. Those two things are almost never the same moment.

Find the peak. Put the right person's name on the ask. Automate the delivery so it happens reliably, every time, without anyone having to remember. That's the whole playbook.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint.

We all chase that feeling, don't we? That surge of accomplishment when you tick off every item on your to-do list, the satisfaction of a perfectly organized inbox, the quiet hum of a day well-spent. But what if some of the very strategies we cling to for productivity are actually slowing us down? What if the pursuit of efficiency has us overlooking something far more valuable: our own unique rhythm and what truly works for us?

I remember a time when I was deep in the trenches of building Asian Efficiency. Every minute felt like it counted, and I was constantly on the hunt for the next big productivity hack. I'd read all the books, listen to all the podcasts, and try to implement every shiny new method. From the Eisenhower Matrix to the two-minute rule, I tried to fit my work into every popular framework. And for a while, it felt like I was making progress. But then, something would inevitably feel off. I'd find myself stressed, trying to force a square peg into a round hole, and wondering why these supposedly universal hacks weren't delivering the promised magic.

It turns out, I wasn't alone in this feeling. My co-host, Brooks, and I recently stumbled upon a fascinating discussion online where people were openly questioning some of the most revered productivity advice out there. It sparked a conversation that made us realize something crucial: true productivity isn't about blindly following rules, but about understanding yourself and adapting strategies to fit your life, not the other way around.

Beyond the Buzzwords: Rethinking Relaxation and Recovery

One of the most intriguing concepts we discussed was “Type A Relaxation.” This is for those ambitious overachievers who struggle to just chill out. They need their relaxation to be an activity they can excel at, like NSDR (non-sleep deep rest), float tanks, or cold plunges. While these can be beneficial, the core idea is to reframe idle time as good, not bad. It's about proactive recovery, like a walk along the water, rather than just passive rest. The key is to cultivate positive coping mechanisms for stress, turning potential setbacks into opportunities for wellness.

Overrated Productivity Hacks: A Reality Check

We dove into a Reddit thread that called out some popular productivity hacks as overrated. Here's what we found:

1. Eisenhower Matrix: While helpful for triage, it's not a daily driver. The delegation aspect often gets a bad rap, but it's a long-term play. Investing time upfront to train someone can free up immense time later. It's about context… what works for a solo entrepreneur might not work for a large team.
2. The Two-Minute Rule: This one often gets misunderstood. It's not about a strict two-minute timer, but the spirit of taking immediate action on small tasks. It builds momentum and helps you get better at estimating. I use it all the time for email, and it's a game-changer.
3. Seinfeld Strategy (Don't Break the Chain): Some say it relies on motivation, but I think it's the opposite. It offloads motivation to an external visual cue. It's fantastic for building daily habits, like journaling or exercising. And if you miss a day? No big deal. Just don't miss twice in a row.
4. Zero-Based Calendar (Hyper-scheduling): This one Brooks and I agree on. It's too much overhead. Life throws curveballs. Focus on outcomes for your day, not scheduling every minute. Buffer time at the end of the day can be a lifesaver for catching up and staying organized.

And the Pomodoro Technique? Many on Reddit called it overrated, but I still use it daily. If you can't focus for more than five minutes, the problem isn't Pomodoro… it's something deeper that needs addressing.

The Four-Day Work Week: A Glimpse into the Future?

Remember when we talked about the UK's four-day work week experiment back in 2022? Well, the results are in, and they're pretty compelling. The majority of companies that participated made the policy permanent. Work intensity remained lower, job satisfaction was higher, and employees felt they performed better. It's not just about employee happiness; it's about business results.

This isn't a one-size-fits-all solution, especially for startups. But for established organizations with solid processes, it can be incredibly effective. The Netherlands even has laws giving employees the choice for a four-day week. It's a fascinating shift, and it makes you wonder if other countries will follow suit.

Your Action Item: Question Everything

So, what's the takeaway from all this? It's simple: question your productivity methods. If something isn't working for you, don't just stick with it because it's popular. Is that technique truly serving you? Or are you just doing it because you feel like you should? Take a moment today to reflect on your own habits and strategies. You might just discover a hidden path to greater productivity and peace of mind.

I asked why some of them never got produced.

“Manufacturing is expensive,” she said. “And I never know which ones will actually sell.”

That’s a real constraint. You design something, have it made, list it — and then find out whether people want it. The risk is front-loaded. If the piece doesn’t sell, you’ve already spent the money.

I showed her a different way to think about it.

The Demo

AI image generation has gotten good enough that you can take a design concept — even a rough sketch, even a written description — and produce something that looks like a professional product photo. Not an obvious render. Not a drawing. Something that looks like the real object, shot under good lighting, on a clean background.

I’ve done this with architecture: a client had hand-drawn sketches of spaces they were trying to sell, and we turned them into photorealistic room renders that looked like finished design photography. The prospective buyers responded to the renders completely differently from the sketches — because the renders felt real.

The same logic applies to any product category.

For jewelry, you start with a description or a sketch. You feed it into Gemini with some reference images for style and scale. You iterate until you have something that looks like a real piece, photographed on a hand or against a clean surface. Then you take that image and put it to work.

The Business Model Shift

Here’s where it gets interesting from a business standpoint.

Traditional product development: design it → manufacture it → list it → find out if it sells.

Risk is at step two. Once you’ve manufactured, you’re committed.

AI-enabled product development: design it → generate a photorealistic image → list it on a pre-order or waitlist page → find out if it sells → then manufacture.

The risk moves to after you have market signal. You don’t spend money on production until you know people want it.

When I walked the jewelry designer through this, I described it this way: imagine you post the image, run a little traffic to it, and collect pre-orders. If you get fifty people who want this piece — make it. If you get three — that tells you something. Move on to the next design.

She immediately realized something: she had a full drawer of designs she could test this way. Not sequentially, not expensively — by putting real-looking images in front of real customers and watching what happened.

Why This Works Beyond Jewelry

The principle applies across product categories:

Apparel. Design a jacket, generate a model shot in that jacket, run a pre-order page, cut the pattern only if you hit a threshold.

Home goods. A candle line, a ceramic piece, a textile design — all of these can be visualized realistically before the first prototype is made.

Food and specialty products. Package design, label design, new flavor concepts — you can test the visual and the concept before production.

Accessories, gadgets, niche hardware. Anything where tooling or initial production runs are expensive enough that a wrong bet hurts.

The constraint used to be that you needed a real product to photograph. Now you need a clear enough concept for the image tools to work from. That’s a fundamentally different starting point.

How to Actually Do This

The workflow isn’t complicated, but there are a few things that matter for getting images that look genuinely professional rather than obviously AI-generated.

Start with good reference images. Give the model examples of the style you’re going for — other products in your category that have the look and feel you want. The more visual context you provide, the more accurately it can match your intent.

Be specific about materials and settings. “A gold ring with a small diamond” produces generic results. “A 14k yellow gold band with a 0.3ct round brilliant diamond, shot on a marble surface, natural window light from the left” gives the model enough to work from.

Iterate in fresh sessions. Image quality degrades in long chat threads as the model loses track of what you established at the start. When you’ve found a direction that’s working, start a new session and rebuild from your best prompts.

Use the output as a proof of concept, not a final asset. The pre-order page image doesn’t need to be perfect — it needs to be clear enough that a real buyer would understand what they’re ordering and feel confident about it. You can invest in proper photography once you have confirmed demand.

The Bigger Shift

What’s actually changed here is where the risk sits in the product development process.

For most of the history of making things, you had to commit capital before you had signal. Manufacturing required it. Distribution required it. The businesses that survived were the ones that had good instincts about demand, or enough margin to absorb mistakes.

AI images change the cost of getting signal. Generating a photorealistic product image is cheap. Running a pre-order page is cheap. Finding out whether fifty people want a thing before you make it — that used to be expensive. Now it’s fast and nearly free.

The jewelry designer had a drawer full of designs she’d been sitting on for years. Some of them probably would have sold. Some of them probably wouldn’t. She had no cheap way to tell the difference.

Now she does.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint — designed to take people from occasional AI use to real, running workflows.

I started with a question.

“How many of you have transformation photos you can’t post? The client’s hair looks incredible, the work is some of your best, but there’s hair on the floor, electrical cables running through the frame, and clutter in the background. Something that makes it unusable.”

Every hand in the room went up.

“How many of you have hundreds of photos like that?”

Most hands stayed up.

“Thousands?”

A few people laughed. Still hands up.

This is the photo graveyard. Every service business has one.

What Happens in the Graveyard

Think about what that actually represents. A stylist spends two hours on a transformation. The client is thrilled. The result is genuinely stunning. Someone pulls out a phone to capture it — and then later, sorting through the camera roll, sees the hair sweep that didn’t happen yet, the charging cable draped across the mirror, the cart of supplies in the corner of the frame.

And the photo goes into the graveyard. Never posted. Never shown to a potential client. Never added to the portfolio. Just… sitting there.

Multiply that by every stylist in a salon, every week, for years. One person at my session said she had over 500 photos like that. Another estimated thousands.

That’s not just a missed social media opportunity. That’s evidence of real work that nobody ever sees.

The 20-Second Demo

After the show of hands, I asked if anyone wanted to try something live.

We took one of those photos — a real one, pulled from someone’s phone in the room. Uploaded it to Gemini. I gave it a simple prompt: remove the background clutter, clean up the cables, clear the floor.

Twenty seconds.

The cables were gone. The hair on the floor was gone. The background looked like the photo had been taken in a clean, properly staged environment.

The room went quiet. Not the polite quiet of an audience following along — the quiet that happens when something genuinely surprises people.

Then someone said, “I have 500 of those.”

What had been a graveyard for months became a content library in under a minute.

This Isn’t a Salon Problem

I want to be clear about something: the salon context made this demo vivid. But the underlying problem is everywhere.

Personal trainers take photos in gyms that have other clients in the background, equipment mid-use, awkward lighting from windows. Contractors photograph finished work with construction debris still in frame. Restaurant owners snap photos of new dishes with the setup clutter not yet cleared. Event photographers capture beautiful moments with venue chaos behind them.

Anyone who takes photos of real work, in real environments, has a version of this problem. The work is good. The background isn’t ready. The photo goes unused.

The solution is the same in every case: a few seconds in an AI image tool, a simple prompt, and a photo that was previously unusable becomes shareable.

Why This Demo Works So Well

After three standing-room-only sessions at that conference, I had time to think about why the photo demo landed so consistently.

It’s not because it’s the most powerful AI use case. It isn’t. It’s because it solved a problem people already had, using work they had already done.

Most AI demos show something futuristic. A workflow that sounds impressive but requires you to build something new, change a process, or learn a new tool from scratch. The gap between “I could see how that would help” and actually using it is wide.

The photo demo has almost no gap. You have the photos already. You open Gemini. You upload. You get the result. The only thing that changed is now you know this is possible.

That’s the pattern I keep seeing with AI adoption in non-technical industries. The tools don’t need to be complicated. They need to be specific. A salon owner doesn’t need to know what a language model is — she needs to know she can fix that photo from March that’s been sitting unused.

What to Actually Do With This

If you have a graveyard of photos like this — from any industry, any context — here’s the workflow:

First, sort through your camera roll and pull out the photos that show good work but have problematic backgrounds. Don’t overthink it. If you’ve been putting off posting something because of the background, that’s the one.

Then open Gemini (free to use, no setup required) and upload the photo. Ask it to remove specific elements — cables, clutter on the floor, items in the background. Be specific about what’s in the way.

Review the result. Gemini is good at this, not perfect. Some photos will need a second pass or a more specific prompt. But most of the time, 20 seconds is accurate.

Start posting the ones that come out well. Build the habit. Go back to the graveyard periodically and work through more of them.

The goal isn’t to use AI for every photo. It’s to stop letting great work sit unused because the background wasn’t clean.

The Bigger Pattern

What struck me most about those sessions was the hands going up before the demo.

Everyone already knew they had this problem. Nobody had thought of AI as the solution, because nobody had shown them a solution that worked for their specific situation. Once they saw it — really saw it, live, on a real photo — the barrier disappeared.

That’s what I think is still underestimated about where we are with AI adoption. The tools have gotten genuinely good at specific, practical tasks. The gap isn’t capability. It’s translation — helping people see where the tool connects to the work they’re already doing.

The graveyard of unpublishable photos is just one example. But it’s a good one, because the before is so clear and the after is so immediate.

Check your camera roll. You probably have a version of this.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint — designed to take people from occasional AI use to real, running workflows.

“I’m basically a glorified typing monkey at this point.”

He looked confused. I was describing my actual workflow for building software — one where I write the spec, describe the vision, and then two AI agents handle the rest. Claude Code builds. Codex reviews. I approve merges. That’s my job now.

He laughed. Then he said, “That’s kind of incredible.”

Yeah. It kind of is.

The Actual Workflow

Here’s how it works in practice.

I open Warp — that’s my terminal — and I pull up two panes side by side. Command-D gives you a vertical split. On the left, I launch Claude Code. On the right, I launch Codex.

These are two separate coding agents. Claude Code is Anthropic’s. Codex is OpenAI’s. Different models, different strengths, and critically — a different perspective on the same code.

I give Claude Code a spec. Not a vague request — a proper spec with the outcome I want, the features needed, any constraints or preferences. Then I say: make a plan, then build it.

Claude Code does its thing. When it’s done with a feature, it commits the code and creates a pull request. That’s the moment it hands off.

Why Codex Reviews Claude Code’s Work

Here’s the thing most people don’t know about AI coding agents: they’re genuinely bad at reviewing their own work.

Claude Code will produce code, look at it, and say it looks great — because it has no fresh perspective on what it just wrote. It can’t step back and see what’s missing or where the logic breaks down.

Codex can. Different model, different training, different set of eyes on the same PR.

So when the pull request is created, I switch to the right pane and tell Codex: “Review PR number X. Fix any issues if needed. Leave a comment.”

Codex pulls up the diff, reads the code, tests it. If something’s off, it fixes it. If everything looks good, it leaves a review comment saying so. Then Claude Code comes back, reads the comment, and merges it.

That’s the loop. Build → review → merge. Two agents. No me in the middle.

Running It Overnight

I usually have 3-4 instances of this running simultaneously on separate branches. Each pair is working on a different feature. Because they’re on separate branches, they don’t conflict — GitHub keeps everything isolated until it’s time to merge.

Sometimes I leave this running overnight. Write the specs before bed. Set it going. Go to sleep. Wake up and check GitHub.

New features. Merged. Sometimes a whole app, depending on how ambitious the spec was.

The quality depends almost entirely on how good the spec is. Really good spec? The agents can get to 80-90% of what I wanted without a single interaction. Vague spec? You get vague results. The writing is now the work.

What I Actually Do All Day

What changed is where my time goes.

Writing code: almost zero. Running things manually: barely. Debugging random errors by hand: rarely.

What I do is: write specs. Review PRs. Say approve. Sometimes redirect.

That sounds like nothing. But it’s actually the hardest part — knowing what to build, in what order, with what constraints, to what quality bar. That thinking doesn’t get automated. It gets more important.

The concept at play here is what I think of as multi-tool native — the idea that these models aren’t interchangeable. Claude Code is great at building. Codex is great at reviewing. You get the best of both by routing the right work to the right model, not picking one and using it for everything.

The Foundation You Still Need

I told Jacob — who had never used GitHub before this session — that this workflow only works once you understand what’s underneath.

Branches. Commits. Pull requests. Reviews. Merges.

Not so you can do all of it yourself. You won’t. The agents will. But you need to speak the language so you can orchestrate them correctly. So when Claude Code says “should I push to main or create a feature branch?” you know what that means and why it matters.

Jacob spent 8 months watching YouTube videos trying to learn to code. He barely got anywhere. Then we started doing hands-on sessions — actually building things, actually using the tools — and six weeks in he’s already at an intermediate level.

The difference isn’t intelligence. It’s method. Doing beats watching. Every time.

The Bottom Line

I’m a glorified typing monkey. I write specs and say approve. My agents ship code around the clock.

That’s the 2026 version of building software.

The people who figure out how to orchestrate these tools — who learn the underlying mechanics, write strong specs, and stay in the judgment seat — they’re going to build things at a pace that wasn’t possible two years ago.

The tools are ready. The question is whether you are.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint — designed to take people from occasional AI use to real, running workflows.

The other 38 run lean.

This is not an accident. It’s how I think about building a stack of agents that actually works — not just technically, but practically. Over time. At scale.

Here’s the framework I’ve landed on.

Meet Teddy and Veto

My two most important agents have names.

Teddy is my executive assistant. He handles scheduling coordination, inbox triage, meeting prep, and a dozen other things that used to eat my mornings. Veto is my task backlog clearer — he processes the queue of things I’ve accumulated and helps me decide what to actually do about each one.

Both of these agents run every single day. Both of them touch work that is high-stakes and deeply personal. If either one messes up, I feel it immediately.

So when I built them, I gave them everything. A 20-page context profile: how I think, how I communicate, what my businesses are doing, what my priorities are, who the key people in my life are, how I make decisions. The same document I use to onboard a real assistant, but this one gets loaded into an AI agent.

Loading that document costs tokens. But for Teddy and Veto, it’s completely worth it.

Why the Other 38 Run Lean

The natural question: if context makes agents smarter, why not give it to everyone?

Simple. Context costs tokens. Tokens cost money. And most agents just don’t need it.

My email sorter doesn’t need to know my 5-year plan. My social media scheduler doesn’t need my sleep patterns. My invoice processor doesn’t need my communication philosophy. Each of these agents has one job. Clear, specific instructions are enough. Loading a 20-page profile into them would be expensive and mostly irrelevant.

There’s also a more subtle reason: when an agent gets too much irrelevant context, it sometimes starts making judgment calls it shouldn’t. Specialized agents work better when they stay specialized.

The concept at play here is what I’d call centralized context — the idea that you do want shared, durable memory in your agent stack, but you have to be deliberate about which agents get access to which parts of it. Not every agent needs the same truth.

The Two Mistakes Most People Make

When I talk to people who are starting to build agent stacks, I usually see one of two patterns.

The first: they give everything to everyone. They load their full context profile into every agent they build, because more context seems like it should be better. The result is high token costs and agents that sometimes behave strangely because they’re reasoning from irrelevant information.

The second: they give nothing to anyone. Their agents run with no persistent context at all. Every conversation starts from zero. The agents feel generic, impersonal, and keep asking the same questions over and over.

The better path is just prioritization. Figure out which agents carry the most weight in your day — the ones where a mistake is costly and where personalization actually matters. Build those out properly. Give them memory. Give them context. Give them everything.

For everything else, keep it simple.

This is basically the 80-20 principle applied to agent building: invest the most in the agents you rely on most, and keep your specialized, single-task agents lean and fast.

One More Layer: The Monitor Agents

Here’s something I added that most people don’t think about until they need it.

Both Teddy and Veto have a monitor agent watching them. A lightweight agent whose only job is to check whether the primary agent completed its task. If something fails — for any reason — the monitor agent immediately sends me a Slack message.

I’ll be honest: I’m a bit of a control freak about this. When your executive assistant is handling mission-critical work, you need to know immediately if something breaks. You can’t find out three hours later when you’re wondering why no one responded to your morning emails.

The setup is simple. In Lindy, you create a second agent with a trigger that fires when the primary agent’s task doesn’t complete as expected. You connect it to Slack (or email, or SMS — whatever you’d actually see immediately). Takes maybe 15 minutes to set up.

The peace of mind is real. I know my stack is running because I’d hear about it if it wasn’t.

How to Apply This to Your Own Stack

If you’re building agents or thinking about building them, here’s the mental model I’d suggest:

First, list every agent you have or want to build. For each one, ask two questions: How often does it run? How bad is it if it messes up?

The agents that run every day and where mistakes matter — those are your Teddy and Veto. Invest in those. Give them context. Give them memory. Set up monitoring.

Everything else? Start simple. You can always add context later. It’s much easier to add than to clean up a mess caused by an agent that was reasoning from irrelevant information.

One last thing: name your important agents. It sounds silly, but it changes how you think about them. Teddy and Veto feel like parts of my team. That makes me more likely to invest in building them well, and more likely to notice when something feels off about how they’re working.

It’s a small thing. But so is the difference between a generic AI chatbot and an agent that actually knows you.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint — designed to take people from occasional AI use to real, running workflows.

Before Google, we navigated the internet through AltaVista. Lycos. Yahoo’s directory. If you were really tech-savvy, you might have had a GeoCities page. Search was clunky. Finding information was hit or miss. The web was useful but not yet transformative for most people.

Then Google launched in 1997 and 1998. And within a few years, the entire landscape of how we worked and found information had changed.

The people who started using Google early — who figured it out while everyone else was still on AltaVista — had a meaningful advantage. They were faster, better-informed, and more capable of navigating the information landscape. That advantage lasted for years.

AltaVista never recovered. Lycos faded. The tools that didn’t adapt became irrelevant.

I think about AI the same way. And I think we’re currently in the 1996-97 moment.

What “Early” Actually Means Here

When I say we’re in 1996-97, I want to be precise about what that means — because there are two mistakes people make with this analogy.

The first mistake is thinking they’re too early. “AI is still developing, I’ll wait until it matures.” That’s like waiting until 2005 to start learning Google. By then, the advantage had already compressed. The early movers had already pulled ahead.

The second mistake is thinking they’re already too late. “Everyone is already doing AI.” No, they’re not. When I work with business owners outside the tech bubble — smart, successful people running real companies — many of them are barely using these tools. The adoption gap is still enormous.

The people learning AI seriously right now are not too early. They’re not too late. They’re in exactly the right window.

What AltaVista Mode Looks Like Today

AltaVista mode with AI looks like this: you’ve tried ChatGPT a few times. You asked it something, got an answer, maybe thought “huh, that’s neat.” And then you went back to doing things the way you always did.

That’s the equivalent of using a search engine for one query and then going back to the phonebook.

There’s nothing wrong with it. AltaVista was useful. But it wasn’t Google.

The shift from AltaVista mode to Google mode with AI looks like: building workflows instead of one-off prompts. Integrating AI into how you work every day, not just occasionally. Understanding what each tool is actually good at and routing work accordingly. Setting up agents that run automatically so you come back to work already done.

That’s the difference between using AI like a search box and using it like a platform.

The Demonstration That Stays With Me

A few months ago, I was showing a finance and accounting team at a real estate company some AI tools they hadn’t seen yet. We walked through how what used to take hours — manual data entry, calculations, reconciliation — could be compressed dramatically.

I did the demo two weeks before the tools even officially launched.

When they actually became available, the finance team went what I can only describe as “crazy” for them. They had seen what was possible. They knew what the tools could do. They were ready.

That readiness — that early exposure — made a real difference. They weren’t starting from scratch when everyone else started. They had a head start.

That’s the Google moment in action.

How Long the Window Stays Open

The Google transition happened faster than most people expected. Within three or four years of Google’s launch, not knowing how to use it effectively was a real disadvantage. The early-adopter window compressed quickly once the mainstream caught up.

I think the same dynamic is in motion with AI fluency — and possibly faster, because the tools are improving at a pace that creates urgency.

“I’ll learn AI later” is going to sound, in hindsight, a lot like “I was going to start using Google eventually.”

Later is fine for some things. For tools that create this kind of compounding, structural advantage — tools where getting fluent early means you’re ahead while others are catching up — later tends to be more expensive than it looks.

What to Do If You’re in AltaVista Mode

The good news about 1997 is that it’s not 2010. Most of the advantage is still in front of you, not behind.

If you’re in AltaVista mode right now — occasional ChatGPT use, no real workflows, no agents running — the first step is just moving to consistent use. Pick one thing you do every day and start doing it with AI involvement every time. Build the muscle.

The next step is moving from prompts to workflows. Instead of one-off questions, build repeatable systems. Meeting follow-ups, content research, proposal drafting, CRM updates — whatever takes your time that’s predictable and repeatable.

Then: agents. Things that run without you. That’s where the compounding starts.

The window is open. The tools are accessible. This is the 1996-97 moment.

The question is whether you’re going to be a Google user or an AltaVista user on the other side of it.

Thanh Pham is the founder of Asian Efficiency. He teaches AI fluency through workshops and the 4-Day AI Sprint — designed to take people from occasional AI use to real, running workflows.

The conversation is fresh. You remember what was decided, what you committed to, what the other person cares about. If you sit down and write the follow-up email right now — in the next 10 minutes — it flows naturally and it lands well.

Most people do not write it in the next 10 minutes.

They jump to the next call. They deal with a fire. They tell themselves they will get to it this afternoon, and then this afternoon becomes tomorrow, and by that point the conversation has gotten cold.

This is one of the most common productivity gaps I see — not in the meetings themselves, but in what happens after them.

What Most AI Notetakers Actually Do

The first wave of AI notetaking tools solved a real problem. Fireflies, Otter, Granola — they capture the conversation, transcribe it, clean it up, and send you a summary. Some of them do this well.

But they stop there.

The summary is useful if you need to remember what was said. What it does not do is take the next action for you. You still have to read the summary, figure out what the follow-up should say, open your email, write it, and send it.

That is five steps that most people defer.

What Lindy Does Differently

I have been using Lindy notetaker, and the difference in behavior it creates is pretty significant.

When a meeting ends, Lindy does not just send me a summary. Within about 30 seconds, there is a pre-drafted follow-up email waiting in my inbox. It references what we talked about, captures any commitments I made or the other person made, and proposes next steps.

I did not write it. I did not instruct it in the moment. It just happened, automatically, as soon as the meeting ended.

The email is usually about 90% right. I do a quick read, make a light edit if needed, and send. Total time: under two minutes.

Compare that to the alternative: remember to write the follow-up, carve out time to do it, reconstruct the context from memory or a summary, write it, edit it. That is 15-20 minutes minimum, and it happens hours later when the window has already closed.

The Sales Team That Fixed Their Follow-Up Problem

I worked with a sales team that had a version of this problem at scale.

They were good at the meetings themselves. But the post-meeting admin was a consistent failure point. The CRM was not being updated consistently. Action items were slipping. Follow-up emails were going out two or three days after calls — if they went out at all.

We built an agent that triggered automatically when a call recording was available. It would update the CRM with key details from the conversation, create tasks for any action items, and draft a follow-up email ready to send.

What used to be 30 minutes of post-call admin work became zero. The rep finishes the call, the system handles everything, and the only thing left is a quick review before hitting send.

That is what happens when you treat the meeting transcript as an asset rather than a record.

A Bonus: The Decision Log

While on the topic of getting more out of meeting transcripts, there is one more thing I have added to my Lindy summary prompt that not many people talk about.

It is a decision log.

Most meeting summaries capture what was discussed and what the action items are. But they do not always make it clear what was actually decided. Those are different things.

I added a section to my summary prompt that asks Lindy to extract every decision that was made in the meeting — not just what was discussed, but what was agreed to.

The use case: if someone missed the meeting or joined late, a decision log tells them exactly what they need to know to stay aligned. Not a full summary — just: here is what was decided, by whom, and what it means.

For any team that runs regular executive or strategy meetings, this is a small prompt tweak that pays off consistently.

The Broader Point

Most of the leverage in meeting automation is not about capturing what happened. It is about what you do with it in the 30 minutes after the meeting ends.

If your notetaker is just storing the transcript, you are leaving most of the value on the table.

The follow-up email, the CRM update, the task creation, the decision log — all of that can be automated off the same transcript. You just need a system configured to do it.

Lindy handles this well if you set it up right. The 4-Day AI Sprint covers meeting intelligence workflows if you want to build this out in more depth.

I will be talking to a smart, successful business owner — someone running a real company with real revenue — and I will ask them how they are using AI. And often the answer is: not much. Maybe they have tried ChatGPT a few times. They have heard a lot about it. But they have not built anything with it. They are not sure where to start.

This surprised me at first. I assumed everyone was further along. When you spend time in tech circles, you are surrounded by people building agents and debating model benchmarks and sharing prompting techniques. It is easy to assume that represents the whole market.

It does not. Not even close.

The Access Problem Is Solved

ChatGPT is free. Claude is a browser tab away. Gemini is built directly into Google products. Perplexity is free for most searches. The tools are available to anyone with an internet connection, at essentially zero cost.

There is no access problem. That gap closed fast.

The gap that exists now — the one that actually matters — is skill.

Skill meaning: knowing how to use these tools to get real results. Knowing how to write prompts that produce useful outputs. Knowing how to connect tools into workflows so things happen automatically. Knowing when to use which model for what task. Knowing how to build an agent that runs without you.

That is a real gap. And it is not closing at the same rate as access.

The Demand Is Real

Earlier this year I started running AI workshops in Austin. I was not sure what to expect — I put together a session, invited some contacts, and sold tickets.

It sold out immediately.

I ran a second one. Sold out again. Five of the attendees from the first workshop came back for the second one, just to stay current on what had changed.

That is not what happens when a topic is saturated. That is what happens when there is a real gap between what people know and what they need to know.

Why the Power Law Matters

Here is the part that does not get enough attention.

The distribution of results from AI is not a bell curve. It is closer to a power law.

A bell curve would mean: people with strong AI skills are maybe 20-30% more productive than people with average AI skills. That is meaningful but not alarming.

What I actually see is more extreme. The people who have really figured out AI — who have moved through the progression from basic prompting to workflows to building autonomous agents — have a highly concentrated amount of leverage compared to people who have not. They are not 20% ahead. They are operating at a different scale entirely.

This makes sense when you think about it. The benefits of AI compound. If you are using AI to generate content, research, and analysis, and you are also using agents to automate repetitive work, and those agents are saving you 239 hours per week — that is not an incremental advantage. That is a structural one.

And structural advantages are hard to close once they are established.

The Three Levels — Where Are You?

When I work with people on this, I use a simple three-level framework to figure out where they are and what the next step looks like.

Level one is AI Assisted. You are using AI tools for individual tasks. You prompt, you get a result, you use it. This is where most people are. It is useful — probably a 1.5-2x multiplier on certain tasks — but it does not compound.

Level two is AI Workflows. You have started connecting tools and building repeatable processes. Instead of one-off prompts, you have systems: a research workflow, a content workflow, a meeting follow-up workflow. This is where things start to get interesting. You are not just saving time on individual tasks — you are removing whole categories of work.

Level three is Building Agents. You are deploying AI that runs without you. You define what needs to happen, set up the system, and come back to finished outputs. This is where the big leverage numbers come from.

Most people are at level one. The gap between level one and level two is where the most accessible productivity gains are. And most people have not crossed it yet.

What to Do About It

The access problem is solved. You do not need to wait for better tools or lower prices. Everything you need to meaningfully move up this progression is available right now.

What is needed is the investment in skill. That means time spent learning, building, and iterating — not just reading about AI but actually using it to build something real.

The window where early adopters have a meaningful advantage does not stay open forever. In most technology curves, early movers get the biggest gains, and then those gains compress as the skill becomes more common.

We are still in the early part of that curve for AI fluency — particularly at levels two and three. The gap is real. The tools are available. The question is whether you are investing in closing it.

When he is developing a new film idea, he has a brainstorming process that surprised me when he described it.

He takes the idea to ChatGPT first. Types out the concept, has a conversation, sees where it goes. Then he copies everything — the original idea, the back-and-forth, what came out of it — and pastes it into Claude. Same question, new conversation. Then takes that and brings it to Gemini.

Three separate sessions. Three different models. All on the same idea.

When I asked why, he was pretty direct about it: every model has different ideas and insights that the other models could not see.

He had found through experimentation that ChatGPT would catch certain angles, Claude would push the concept in directions ChatGPT had missed entirely, and Gemini would surface things neither of the first two had touched. Deliberately running the same idea through all three gave him a more complete picture than any single model could.

Why This Works

The AI models available today are genuinely different — not just in capabilities, but in how they approach problems.

ChatGPT tends to be strong on general brainstorming and broad ideation. It has been trained on a huge range of content and has a natural fluency in exploring possibilities. Claude tends to have a more analytical bent — it often pushes back, asks clarifying questions, and looks for structural coherence. Gemini draws on Google data and tends to surface connections that the other models might not reach.

For a film producer brainstorming story ideas, those differences matter. A good story idea needs to hold up to different kinds of pressure — emotional resonance, logical coherence, market relevance. Running the idea through three different editorial minds is a fast way to stress-test it from multiple angles.

This is what I think of as multi-tool native thinking. The best AI users do not pick one model and stay loyal to it for everything. They learn what each tool is actually good at, and route their work accordingly.

The Manual-to-Automated Arc

When he described his process to me, I saw two things simultaneously.

The first was how smart his underlying logic was.

The second was how much friction there was in executing it. Copying and pasting between three separate browser windows, reformatting the context each time, keeping track of what each model had said — it was a lot of overhead for something that could be systematic.

So I built him a workflow. He puts the idea in once. The workflow runs it through all three models, collects the responses, and returns three distinct perspectives in a single output.

Same result. One step instead of ten.

That is usually how automation decisions should work: you identify something someone is already doing because it genuinely adds value, then you remove the friction around doing it. You do not automate things people are not doing for good reason. You automate things people are already doing manually for good reason.

How to Apply This Without Building a Workflow

You do not need a custom workflow to try this. You can do the manual version yourself, and it is worth doing at least once to see if it changes anything for you.

Pick a problem you are working through — a strategic decision, a creative concept, a piece of writing you want to get right.

Take it to ChatGPT first. Have a real conversation, not just a single prompt. Then copy the full context and bring it to Claude. See what is different. Then take both to Gemini.

You will notice that the models disagree in interesting ways, emphasize different things, and sometimes surface angles you had not considered. That is the point.

Once you have done it manually a few times and confirmed it is actually useful for your work, then it makes sense to build or find a workflow that reduces the friction.

The principle behind what the film producer was doing is simple: for complex problems, multiple perspectives beat one. That is true whether you are talking to people or to AI models. The best ideas usually get sharper when they are pressure-tested from different angles.

AI just makes getting multiple perspectives faster than it used to be — if you know to ask for them.

Read the message. Switch to ChatGPT. Type out some context. Type out what I wanted to say. Edit it a little. Copy the output. Paste it into my email. Edit it again. Send.

Faster than writing from scratch? Yes. But it was still four or five steps and 10-15 minutes per email.

Today, I read the message, talk out my reply, and it is 90% ready to send.

The change: I stopped typing.

The App That Changed It

The tool is called WhisperFlow, and it is my highest-leverage single app right now.

Most people have tried their phone or Mac built-in voice dictation and found it frustrating. You have to speak in perfect sentences. If you pause or stumble, it produces garbled text. The output needs heavy editing. So they give up and go back to typing.

WhisperFlow is different because there is an AI processing layer between what you say and what appears on screen.

You can ramble. Think out loud. Start a sentence, change direction, come back to it. The AI cleans it up into clear, properly formatted text. You are not transcribing speech — you are letting your natural thinking flow, and the AI turns it into something usable.

That one difference makes voice dictation actually practical.

The Speed Math

A detailed prompt that used to take me 20 minutes to type now takes about 5 minutes to dictate.

For emails, it is even more dramatic. I used to spend 10-15 minutes on a substantive email reply — reading, thinking, typing, editing. Now I am under 5 minutes for most emails. Read, talk, light cleanup, send.

Across a full week of emails and prompts, that is hours of time back.

But here is the part that surprised me more than the speed.

Voice Unlocks Different Thinking

When you type, you edit as you go. Every sentence gets filtered before it leaves your fingers. You trim ideas before they are fully formed. The text you produce is clean, but it is also compressed — you lose some of the nuance and context that was in your head.

When you talk, you think out loud. The thinking is less filtered. You say things you might not have typed. You add context you might have left out.

The prompts I dictate tend to be richer than the ones I type. More context, more nuance, more of the actual thought behind the request. And when the prompt is better, the output is better.

This surprised me because I assumed voice dictation was just a speed optimization. It turned out to also change the quality.

How I Use It Day-to-Day

WhisperFlow works across everything — ChatGPT, Claude, email, Slack, anywhere you type text.

For prompting: I open whatever AI tool I am using and instead of typing, I just hit the WhisperFlow shortcut and talk through what I want. Long, detailed prompts that would have taken me 20 minutes now take 5.

For email: I read the incoming message, then talk through my reply out loud. The context is fresh in my head, so the response is usually close to what I would write. Quick cleanup and done.

For messages: Same thing. Whether it is Slack or a text, I dictate more than I type now.

I also built this into a broader tool stack for the days I am moving fast. WhisperFlow handles the input layer. Lindy handles the automations that run in the background. Paste, a clipboard manager, makes it easy to store and reuse screenshots and templates. Together, they have changed the texture of my day more than almost anything else I have done in the last year.

How to Start

WhisperFlow is available on Mac and iPhone. The setup takes a few minutes — you set a keyboard shortcut, and from then on, you press the shortcut wherever you want to dictate.

The learning curve is mostly about getting comfortable talking to your computer. If you have never done it, it feels awkward for the first hour. By day three, you will not want to go back.

Start with email. Pick one reply tomorrow that you would normally spend 10 minutes writing. Dictate it instead. See what the output looks like.

If you are spending more than a couple hours a week typing prompts or composing messages, this is worth 10 minutes to try. It is the single highest-leverage app in my current stack — and I do not say that lightly.

Visit www.asianefficiency.com for more productivity tips and tactics.

Try Notion Custom Agents at notion.com/tps.

Links

• Notion Custom Agents
• 25X Productivity Coaching

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

And I notice the same thing every time.

People come in having used ChatGPT. They know how to prompt. They get value from it. But there is almost always a moment in the workshop — usually when I show an agent that just produces something finished, on its own, without them typing a single thing — where they stop and say: I did not know it could do that.

What they are discovering is the difference between two completely different modes of using AI.

Mode 1: AI Working With You

This is where most people live, and it is useful. You open ChatGPT. You type something. You read the response. You type back. You go back and forth, refining, exploring, building on what it gives you.

This is conversation mode. You are present the whole time. The AI is a thinking partner — smart, fast, available around the clock. For brainstorming, drafting, and research, it is genuinely great.

But here is the thing: you are in the loop for every step. Nothing happens unless you are there. The AI helps you do the work. You are still doing the work.

Mode 2: AI Working For You

The second mode looks different.

You define a task once — what it is, what a good output looks like, what the AI needs to complete it. You set it up. And then it runs without you.

The output is an artifact. Not a conversation — a finished thing. An email sent. A report built. A proposal formatted and ready. A document created. A task logged in your CRM. A summary delivered to your inbox.

You delegated. The AI produced the outcome. You come back to find it done.

This is what it means for AI to do work for you, not with you. In conversation mode, you get help. In delegation mode, you get results.

Why Most People Get Stuck in Mode 1

When I talk to people who have been using AI for a while but have not made this jump, the reason is usually one of two things.

The first is that they do not know mode 2 exists. Nobody showed them. They experienced conversation mode first and assumed that was the whole product. ChatGPT is designed around conversation — it is the default interface, and it shapes how people think about what AI can do.

The second reason is fear. In workshops, I see this clearly. People who are perfectly comfortable generating text in ChatGPT get very cautious the moment an agent might actually send something. An email drafted is fine. An email sent? That is scary. They want to stay in the loop.

That caution is reasonable — especially at first. But it keeps people in a mode where they are always doing more work than they need to.

The Practical Difference in Leverage

In conversation mode, AI scales your thinking. You can explore ideas faster, draft faster, research faster. That is real value — probably a 2-3x multiplier on certain tasks.

In delegation mode, AI scales your output. You are not in the loop, so you can be doing something else entirely while the work gets done. You can have multiple workflows running in parallel. You can wake up to things that got completed overnight.

The leverage ratio is completely different.

One framework I use for figuring out where to build delegation-mode systems: what do I do every day or every week that requires almost no judgment? Those are the tasks worth automating first. Not the impressive one-off things — the repetitive, predictable, time-consuming things that happen over and over. That is where compounding ROI comes from.

How to Start Moving Into Delegation Mode

You do not need to overhaul everything at once.

Pick one task you do every week. Something predictable, something where the output is pretty consistent, something where you have done it so many times you could describe exactly what a good result looks like.

Build one agent for that one task. Run it long enough to trust it. Then build the next one.

The transition from conversation mode to delegation mode happens one workflow at a time. Most people who are running serious agent systems now did not start with 10 agents. They started with one — usually something small, like a meeting summary or an email draft — and kept building.

The question worth sitting with: what in your work right now could produce a finished artifact without you being in the loop?

That is your starting point.

I did not have time to read five books and experiment for months. So I went to Clarity.fm and booked one hour with someone who had already figured it out. He walked me through his whole system. That one hour saved me weeks of trial and error.

The principle was simple: when someone has already solved something and published what they know, the fastest path is to learn directly from that knowledge — not to start from scratch.

NotebookLM takes that same principle and applies it to YouTube.

What NotebookLM Actually Does

NotebookLM is a free research tool from Google that lets you upload sources — documents, PDFs, links — and then ask questions, generate summaries, build mind maps, create quizzes, and produce audio overviews from that material.

What most people do not know is that there is a Chrome plugin that lets you load entire YouTube channels into NotebookLM in one click.

Here is the example I ran in a recent workshop: Lex Fridman channel. Close to 300 videos, most of them two hours or longer. That is around 600 hours of content.

I clicked the plugin. All 300 videos loaded into my NotebookLM notebook. Then I could ask anything across all of it:

• What are the most common themes across his AI interviews?
• Summarize his view on consciousness based on everything he has discussed
• Build a mind map of his recurring frameworks on learning and curiosity
• Create a quiz on his key ideas
• Generate an audio overview to listen to on a walk

600 hours of content. One click. Queryable from a single interface.

The Before and After

I worked with a client who was doing something close to this manually — and it shows why the friction matters.

He was interested in following investment creators on YouTube for ideas. His process: take a YouTube URL, paste the transcript into ChatGPT, write a custom prompt, extract the relevant stock tickers and sector analysis. One video at a time. About 30 minutes per video.

Useful. But with that much friction, he was only covering 5 or 6 creators a week.

We built one Lindy agent to handle the whole flow automatically. He submits a YouTube URL. Within seconds, he gets a structured summary: key ideas, mentioned tickers, sector themes, anything worth flagging.

Because the friction disappeared, the behavior changed. He went from 6 creators a week to monitoring 20+.

That is the real lesson. Removing steps does not just save time. It changes what you are able to do at all.

How to Use It

The setup takes a few minutes.

First, go to notebooklm.google.com and create a free account if you do not have one.

Second, search for the NotebookLM Chrome extension — look for one with strong reviews that specifically supports YouTube channel import.

Third, go to any YouTube channel you want to study. Click the extension button. It will add all the videos as sources in a new NotebookLM notebook.

From there, you can use the chat panel to ask questions, click Generate to build an audio overview or mind map, or use the quiz and flashcard feature to test your retention.

One good starting point: load a channel from someone whose work you have been meaning to dig into. Ask it to summarize the top 10 ideas across all their videos. You will get in an hour what would have taken weeks of watching.

Why This Matters Beyond Just Saving Time

There is a principle I think about when building knowledge systems: treat information as an asset, not a stream.

Most people consume content passively. It plays, they half-watch, it goes into the playlist graveyard. The knowledge never really sticks because there is no active engagement with it.

Loading a creator full body of work into NotebookLM forces a different relationship with that content. You are not just watching — you are building a searchable, queryable version of their thinking. That content becomes something you can return to, dig into, and use.

Every expert you follow has already published what they know. For free. NotebookLM is how you actually absorb it instead of letting it pile up.

The consulting fee version is faster but expensive. The YouTube version is free — it just used to require patience and memory. Now it requires a plugin and one click.

One was figuring out what to study. One was already in school but was thinking about changing direction. And I told both of them: go into software and go into accounting.

One became a software engineer. One became a CPA.

At the time, that was genuinely good advice. Software was hard to build — it required real skills, real training, and there was strong demand because most businesses could not build anything without hiring someone who actually knew how to code. Accounting had a massive moat: years of study, regulatory knowledge, and professional certifications. Both fields had high barriers to entry and strong job security.

Today, I would give them different advice.

Not because those fields are dying. They are not. But because the moat has shifted — and if you do not understand how, you are going to be caught off guard.

What Changed

The CEO of Anthropic recently said something that stopped a lot of people in their tracks: most of the software they build internally at Anthropic is now coded by AI, not by software engineers.

That is not a dig at engineers. Anthropic is still one of the most engineering-intensive companies on the planet. But it tells you something important about where the leverage has moved.

Software that used to cost a company $50,000 and three months of a developer time can now be built in days. The barrier to entry for building software — apps, tools, automations — is collapsing. Which means the moat that used to come from knowing how to code is narrowing.

Same thing is happening in accounting. AI tools are handling analysis, categorization, reconciliation, and reporting at a level that would have taken a trained accountant significant time. The technical skills that built the moat are getting easier to acquire — or skip entirely.

The Moat Has Moved, Not Disappeared

Here is what I want to be clear about: this is not a story about AI replacing people.

I work with a lot of professionals across different industries, and the ones I see thriving right now are not the ones who abandoned their domain expertise. They are the ones who took their domain expertise and layered AI fluency on top of it.

A finance professional who deeply understands financial statements and knows how to query AI tools to find patterns, flag anomalies, or model scenarios — that person is doing the work of three people.

A software developer who can architect systems and use AI to generate, test, and iterate on code faster — that person ships things faster than a whole team did five years ago.

The domain knowledge still matters. In some cases, it matters more than ever — because you need to know when the AI output is wrong. But the skill that sits on top of all of it is how well you can work with AI.

The Three Levels I Teach

When I work with people on this, I use a framework I call AI Fluency Levels.

Level one is AI Assisted — you are using AI tools like ChatGPT for one-off tasks. Prompting, generating, reviewing. This is where most people are.

Level two is AI Workflows — you have started connecting tools and building repeatable processes. Instead of copying and pasting between tools, you have systems that do it for you.

Level three is Building Agents — you are deploying AI agents that run autonomously and handle whole workflows without your involvement. This is where the biggest productivity multipliers live.

Most people stop at level one. The gap between level one and level two is where the moat starts to form.

What I Would Tell My Brothers Now

If I sat down with them today, I would say the same things I said in 2018 — learn the fundamentals deeply, get the credentials, build real expertise in your field.

But I would add one more thing.

Get good at working with AI. Not just using it — working with it. Building with it. Understanding what it is good at and where it needs you. Managing a system where you are directing AI to do the repetitive, automatable parts so you can focus on the judgment calls.

That is the skill I did not know to teach in 2018. It is the one that matters most right now.

The people I watch fall behind are not the ones who do not know their field. They are the ones who know their field but are doing everything manually while others stack AI on top of the same domain knowledge and produce at a completely different speed.

The moat has moved. The question is whether you move with it.

Information is becoming a commodity.

Not eventually. Now. The transition is already happening.

The Erosion of Information Scarcity

For most of human history, information was scarce. Knowing things that other people did not know was genuinely valuable. Expertise meant having accumulated, through years of study and experience, a body of knowledge that was hard to access anywhere else.

Lawyers charged premium rates because understanding the law required years of practice. Consultants commanded high fees because they had synthesized patterns across dozens of engagements that most companies had not seen. Financial advisors, marketing strategists, and HR experts — their value rested partly on information asymmetry. They knew things you did not.

AI has been systematically eroding that asymmetry, and the erosion is accelerating.

Ask an AI about tax law in three different jurisdictions. About construction codes for a specific building type. About the psychological research underpinning a marketing technique. About the failure modes of a particular supply chain structure. The breadth and depth of what these systems know now exceeds what any individual expert can carry in their head.

This does not mean experts become worthless. But it does mean that knowing things is no longer a sustainable standalone moat. The information access premium is going to zero.

What This Means for Information-Based Businesses

If your business model is built on being the person who knows things others do not, this is worth taking seriously.

Newsletter writers whose value is staying current and sharing what they find. Consultants who advise on a specific domain. Coaches who teach frameworks they have assembled. These businesses are not disappearing — but the foundation they are built on is shifting.

When I work with business owners and investors on AI, I keep running into the same thing: people on Twitter think everyone is up to speed on AI. The reality is that most business owners — the ones running real companies with real teams — barely know how to write a basic prompt. The gap between the AI-native community and the real market is enormous.

That gap creates a window. But it is a window based on timing, not information scarcity. It closes as adoption spreads.

The Three-Dot Principle

I think the enduring value moves to people who can connect three dots.

Not two — three. Here is why the distinction matters.

Two-dot combinations are the most common play: take your domain expertise and add AI skills. A lawyer who knows how to use AI tools. A marketer who builds AI workflows. An architect who uses AI rendering. These combinations are valuable right now. But they are visible, replicable, and will become the baseline expectation within a few years.

Three-dot combinations are something different. They create categories that do not yet have obvious names. They produce outputs that require all three components simultaneously — and because fewer people hold all three, they are harder to replicate.

The example I keep coming back to: someone who combined interior design expertise, AI image generation skills, and real estate investment experience. That combination lets them show developers photorealistic renders of buildings that have not been built — the kind of visual that closes presale deals before construction starts. None of those three skills alone produces that outcome. Neither does any two. All three together creates something genuinely new.

What This Looks Like in Practice

I have someone on my team, Brooks, who spent time learning operations, digital marketing, and productivity systems. He did not do this because someone handed him a roadmap. He just kept adding adjacent domains.

Now, when we discuss a marketing campaign, he understands the backend mechanics. When I describe a system, he immediately sees the marketing angle. I do not have to translate between worlds — he holds both simultaneously. Add productivity expertise on top of that and you have someone with a combination that is genuinely rare. Very few people know all three.

That is the pattern. It is not about picking the right career track. It is about deliberately accumulating at the intersection of multiple domains — especially domains that do not obviously go together.

The question I keep asking people: what are the three domains you sit at the intersection of? Not your primary one. Three. What combination, if you owned it, would make you hard to replicate?

The New Leverage

The shift I am describing is not a reason to stop learning or to abandon expertise. It is a reason to think differently about where the edge is.

Raw information is cheap. AI made it that way, and that is not reversing. The leverage is not in knowing more than the model — it is in knowing how to synthesize across domains in ways the model alone will not produce without a specific human pointing it in the right direction.

That synthesis is what I mean by connecting three dots. It requires depth in more than one area, pattern recognition across unexpected domains, and judgment about which combinations create actual value.

AI accelerates the people who have that. It makes the synthesis faster, the outputs higher quality, the iteration tighter. But it does not replace the human who knows which three dots to connect.

That is where the value is going. Not to information hoarders. To cross-domain synthesizers.

Want to develop the kind of AI fluency that compounds with your existing expertise? The 4-Day AI Sprint is a hands-on program for people who want to move from knowing about AI to actually building with it.

My working style. My active projects and priorities. Who I am trying to stay in touch with. My preferences for tone, follow-up, and communication. When I am traveling. What I am trying to avoid having on my plate.

This is not a journal. It is a context profile — a document I have built to give my agents the background information they need to actually make decisions on my behalf. And it is probably the highest-leverage thing I have done to improve how my AI agents perform.

The Stateless Problem

Most AI interactions start from zero.

You open a chat. The agent has no idea who you are, what you care about, what you are working on, or what a good outcome looks like for you specifically. Every request requires you to re-establish the context that should already be known. Every session, you are explaining things the agent should already have internalized.

This is not a flaw — it is the default state. AI models do not retain memory across sessions unless you give them something to remember. And most people never do.

The result is AI that is generically capable but never quite calibrated to you. It is the difference between a new hire who technically knows the job and a trusted assistant who knows your preferences, your patterns, and your priorities without having to ask every time.

A context profile closes that gap.

What a Context Profile Is

A context profile is a readable document — mine runs 20 pages — that captures the information your agents need to work effectively on your behalf.

At minimum, it should cover:

• Who you are and what you do — your role, your business, the kinds of decisions you make
• Current priorities and active projects — what is urgent, what is on the back burner, what is coming up
• Working preferences — how you like to communicate, your tone, your schedule and energy patterns
• Key relationships — who you need to stay in touch with and why, context on your most important contacts
• Upcoming travel or schedule changes — anything that would change how the agent should behave
• What to avoid — decisions you do not want made without you, topics that require your direct attention

I learned this approach from Flow, the CEO of Lindy, who has an even longer version than mine. His take: it is the single highest-leverage thing he has done to improve how his agents perform. Once an agent has this context, it stops being a generic tool and starts being something that actually knows how you work.

What Changes When You Load It

The difference shows up in every interaction.

Here is a concrete example. I set up a personal daily podcast — an AI workflow that pulls articles from my reading backlog each morning, converts them to audio using voice synthesis, and delivers them to me as a podcast episode. It is a useful setup, but what made it powerful was loading my context profile into the workflow.

Instead of just summarizing the articles, the AI now discusses them through my lens. It connects the ideas to projects I am actively working on. It surfaces action items that make sense given my current priorities. Same AI, same articles — completely different output, because the agent knows who it is talking to.

This pattern generalizes. An agent handling email suggestions performs differently when it knows your communication style and your relationship with each sender. An agent prepping you for a meeting performs differently when it knows your history with that person. An agent managing your backlog performs differently when it knows which types of tasks you are trying to reduce.

The context profile does not make agents smarter. It makes them more accurate — more calibrated to you specifically.

The Failure Mode: AI That Keeps Forgetting You

The pattern I see most often is what I call the one-night stand approach to AI. You have a great interaction. The agent does something useful. Then you close the tab and the next time you open it, the agent has forgotten everything. You start from scratch.

No continuity. No memory. The same clarifying questions every session. The same generic outputs that do not quite fit your situation.

This is fixable. But it requires treating your agents the way you would treat a new team member who needs proper onboarding — not just throwing them tasks and hoping they figure out what you need.

The fix is a durable, readable context profile that you load into your most important agents. Not every agent. Just the two or three that touch the highest-stakes parts of your work day.

How to Build Yours

The fastest way to create a context profile is not to write one from scratch. That produces something thin and generic.

The better approach: let AI interview you.

Open ChatGPT. Tell it you want to create a context profile for your AI agents — something that will help them make decisions on your behalf. Ask it to interview you. It will ask questions. Answer them honestly and in detail. After 20-30 minutes, ask it to compile your answers into a structured document.

Then review and edit. You will probably end up with something 5-15 pages long. Save it somewhere accessible — Google Docs works well because many agents can read directly from it.

Then load it into your most important agents. For me, that is my email summarizer, my daily briefing agent, and my meeting prep agent. Those three touch enough of my day that having them properly calibrated makes a noticeable difference.

The Payoff

When it is working well, you notice it in small ways at first.

The agent stops asking you to re-explain your role. It makes reasonable assumptions about tone. It surfaces the right things without being asked. It knows when you are in a different context — traveling, in back-to-back meetings — and adjusts accordingly.

The failure mode is an agent that keeps guessing who you are. The success state is an agent that already knows.

Your context profile is the bridge between those two. Build it once. Update it occasionally. Load it into your best agents.

Want to build smarter AI systems that actually work for you? The 4-Day AI Sprint covers practical frameworks for designing agents, workflows, and context systems that fit how you actually work.

His name is Jacob. He is about to start a construction internship — his first real job. He has been learning AI tools. He wanted to know whether it would actually make a difference.

I told him the honest answer: yes, but not in the way most people think.

It is not about having AI on your resume. It is about what you can actually do with it that the person next to you cannot. And more importantly, it is about whether you can show it.

What Is Happening at the Company Level

When I work with companies on AI implementation, I keep seeing the same pattern.

They are not laying people off across the board. The hiring has not stopped. But the math has changed.

What used to require five junior employees — research, drafting, data entry, summarizing, coordinating — is increasingly handled by AI workflows. One well-built agent does the work that three people used to do. That means companies can hire fewer people, upskill the ones they have, and come out ahead on productivity.

The phrase I keep hearing: we would rather train two strong people to do the work of five than hire the five.

This is not a prediction. I am watching it happen right now in the companies and founders I work with. And it has real implications for anyone entering the workforce, or trying to stay ahead in it.

The Moat Is Real

AI fluency — genuinely knowing how to use AI tools to produce, build, and deliver — is becoming a career differentiator in a way that certifications and courses are not.

Here is the distinction I keep drawing: a LinkedIn badge tells me you completed a course. A built tool tells me you can actually ship something.

That gap matters. Because the companies doing the math I described above are not looking for people who know about AI. They are looking for people who can use AI to produce results they can see and measure.

At SXSW earlier this year, I was talking about exactly this. The people who are pulling ahead in the current market are not necessarily the most credentialed. They are the ones who are AI native — who have built enough things to know what is possible, what is worth automating, and how to get good output from AI tools consistently.

The ones who figured this out two years ago have a significant head start. But it is not too late. The moat is still available. The compounding just starts whenever you start.

What I Told Jacob: Build in Public

When Jacob asked me how to stand out, my answer was not take a course. It was: build something real and show it.

Build in public means picking a problem that actually exists — ideally in your industry — and building an AI tool or workflow to solve it. Then documenting that process and putting it somewhere people can see it.

Not a vague project description on a resume. The actual thing you built, with a clear explanation of the problem it solves and the results it produces.

A resume describes what you have done. A built tool demonstrates what you can do. Those are not the same thing, and employers increasingly know the difference.

In construction, Jacob understands what this looks like on a job site. You do not get hired to manage projects by writing about projects. You get hired because you have done it and have something to show for it. AI is the same.

How to Start

If you are trying to build AI fluency — and the career moat that comes with it — the path is simpler than most people make it.

Start with your own friction. What is annoying in your day-to-day work? What takes too long? What requires repetitive steps that never quite seem worth the time? Those are the problems AI tools solve best.

Build the smallest useful version. You do not need to build a complex system. You need to build something that works and saves you time. A workflow that summarizes your emails. A tool that drafts responses from a template. A script that formats data you would otherwise do by hand.

Document it. Screenshot the before and after. Write two paragraphs explaining what you built and why. Post it somewhere — LinkedIn, GitHub, a personal site, a Notion page. The act of documenting forces clarity, and the act of publishing creates the portfolio.

Repeat. AI fluency compounds. Each thing you build makes the next one faster. Each problem you solve with AI teaches you what is possible in adjacent problems. The gap between someone with fifty reps and someone with five reps is enormous — and it keeps growing.

The Timing Is Good

The honest message I give people in Jacob position: the timing is good, but it will not stay this way forever.

Right now, being genuinely AI-fluent — not just curious, but actually building things — still sets you apart. Most people are still in the I have used ChatGPT a few times category. The people doing real work with AI tools have a visible edge.

That window will narrow as more people catch up. The moat will still exist, but the barrier to entry will be higher. What makes you stand out today might just make you competitive in three years.

The people I watch pulling ahead are the ones building now. Not waiting to feel ready, not waiting for the perfect tool or the right course. Just picking a problem and building.

That is AI fluency. And it is the new career moat.

Want to build real AI fluency fast? The 4-Day AI Sprint is a hands-on program for getting dramatically better results from AI tools — by actually building things, not just learning about them.

I don't know how to code. Never have. I couldn't write a function in Python or JavaScript if you put a gun to my head. I wouldn't know where to start.

And yet I've built more software in the last year than most developers build in five.

An agent that reads 100 texts a day, summarizes them, and suggests replies — two hours to build, and it runs every morning at [8:30] AM. A CPA command center for a client, built live on a coaching call. A full content pipeline that distills transcripts into drafts across four platforms, runs automatically every two hours, and publishes to WordPress and Ontraport without me touching it. CRM tools. Image generation workflows. Meeting summary agents.

All of it without a single line of code written by my hands.

So what's actually going on here?

The Skill Is Not Coding

The mistake most people make is assuming that building software requires knowing how to code. That assumption made sense for decades. It doesn't anymore.

The skill is something else entirely. Two things, actually.

First: having ideas for what to build. Knowing what would save you time, what would eliminate friction, what would make your work better. Having enough clarity about your own problems that you can describe a solution.

Second: being dangerous enough to prompt with precision. Knowing what you want and being able to describe it in enough detail that the AI can actually execute it. Not vague requests. Specific, structured direction.

The analogy I use is driving a car. You don't need to understand how the combustion engine works to drive. You need to know how to steer, when to brake, where you're going, and what a good outcome looks like. The engineering happens under the hood. You handle the thinking.

AI-assisted building works the same way. Claude Code does the engineering — writing the functions, setting up the database schema, handling the API connections, debugging the errors. You bring the vision, the requirements, the judgment about what a good outcome looks like.

What I Told Jacob

I was coaching Jacob recently — a college senior, about to start his first job as a construction intern. He wanted to learn AI tools but was intimidated by the idea of building software. He assumed it required skills he didn't have.

When I told him I can't code, I watched something shift.

Because Jacob already knows how to do this. He just doesn't know he knows.

In construction, he reads blueprints. He describes materials and constraints to the people doing the actual work. He communicates requirements clearly enough that contractors can execute. He spots when something's off before the build goes wrong and becomes expensive to fix.

That's the entire skill set for building with AI.

The ability to describe what you want in specific, structured terms. The judgment to know when the output is right and when it's missing the mark. The clarity to communicate requirements before the work starts. These aren't coding skills. They're thinking skills. Communication skills. Domain knowledge.

Jacob has all of them. He just needed to redirect them from talking to contractors to talking to Claude Code.

Start Small, Build Up

One thing I tell everyone learning to build with AI: you don't need to start with the complex thing. Start with one problem. The smallest version of what you actually need.

The text summarizer I mentioned — I didn't build a full communication management system on day one. I identified one specific problem, described the solution I needed in specific terms, and let Claude build the minimum viable version.

It took two hours. Without AI tools, the same functionality would have required months of development time and cost thousands of dollars. That's not a small difference — it's a different category of what's possible.

Once you prove value on the small version, you expand. Add integrations. Improve the prompts. Build the next thing. Life gets better one agent at a time.

The Gap Is Smaller Than You Think

The gap between where most people are and building real software with AI is not a coding gap. You don't need to learn Python. You don't need a computer science degree.

The gap is learning how to interface with the AI so it does exactly what you want.

That means developing a clearer picture of what you need before you ask for it. Learning to recognize when the output is right and when it needs correction. Getting comfortable with the back-and-forth of describing, reviewing, adjusting, and building.

It's closer to project management than to programming. And most people who think they can't build software already have the underlying skills — they just haven't pointed them at AI yet.

Jacob did. Two weeks after our session, he was building his first tool. He didn't learn to code. He just learned to describe what he wanted precisely enough for Claude to build it.

Want to go deeper on building more effectively with AI? The 4-Day AI Sprint covers practical techniques for getting dramatically better results from AI tools in your daily work.

“Which one should I use — ChatGPT, Gemini, or Claude?”

The answer is not one. It is all three. But for different things.

I was explaining this to Jacob during a coaching session recently. He is a construction intern learning AI tools for the first time. He had been using ChatGPT and wanted to know if he was missing anything.

I told him to think about it the way he thinks about builders.

Some contractors specialize in sports facilities. Some in luxury homes. Some in office buildings. They all build — but they are not interchangeable. You would not hire the stadium guy to design a custom home. The specialist exists for a reason.

AI models work the same way.

ChatGPT: The Daily Driver

ChatGPT is the most general-purpose of the three. Research, brainstorming, writing, thinking through decisions, answering questions on any topic. It is the first tool most people reach for — and for good reason.

Think of it like a subject matter expert you can actually have a conversation with. Someone who knows a lot about almost everything, who gives you a real answer instead of a search results page, and who can switch contexts fast.

For day-to-day tasks — emails, research, explaining concepts, working through ideas — ChatGPT handles it. It is the daily driver.

Gemini: The Visual Specialist

Gemini earns its spot on images and video.

When I need photorealistic renders — for a project pre-sell, a product concept, a design mockup — Gemini is what I use. ChatGPT can generate images too, but there is a quality difference. ChatGPT images have a glossy AI finish to them. Gemini renders look like real photography.

I have seen this used to real effect. One of our coaching clients used Gemini renders to pre-sell a project before a single thing was built. Potential buyers made decisions based on AI-generated images that looked indistinguishable from actual photos.

Gemini also integrates natively into Google ecosystem. If you live in Docs, Sheets, Gmail, or Drive, Gemini connects there in ways the other models do not.

Claude: The Builder

Claude is what I open when something needs to actually be built.

Technical work. Coding. Agent design. Complex structured tasks where the output has to be right, not just plausible. Claude is strong at holding a lot of context and reasoning through it carefully — which matters when you are designing something with many moving parts.

One thing most people do not know: Claude can connect to Google Drive directly. Most other models cannot do that yet. So if your source material lives in Drive — transcripts, documents, research files — Claude can work with it without you copy-pasting everything in.

When I am building an agent system or designing a workflow from scratch, Claude is what I reach for.

The Routing Principle

The failure mode I see most often: someone finds a model they like, gets comfortable with it, and tries to force it to do everything.

That works — until it does not. You end up with images that look AI-generated when you needed photorealistic. You end up with code that almost works. You spend 40 minutes on something that would have taken 10 minutes in the right tool.

The better approach is routing. Instead of asking which AI is best, ask which AI is best for this task.

• ChatGPT — general-purpose work, research, brainstorming, daily conversations
• Gemini — images, video, visual concepts, Google Workspace integration
• Claude — technical builds, coding, agent design, anything complex and structured

Once you internalize this, your output gets noticeably better — not because you got better at prompting, but because you stopped asking a luxury home builder to design a sports stadium.

Jacob got it immediately. Because in construction, this is obvious. The specialists exist for a reason.

It is the same with AI. Route the work to the right tool. Do not be loyal to the model. Be loyal to the outcome.

Want to go deeper on building more effectively with AI? The 4-Day AI Sprint covers practical techniques for getting dramatically better results from AI tools in your daily work.

Visit www.asianefficiency.com for more productivity tips and tactics.

Sign up for a $1/month trial period at shopify.com/tps.

Get the right life insurance for you and save more than fifty percent on term life insurance at SelectQuote.com/TPS.

Try Notion Custom Agents at notion.com/tps.

Masterclass.com/TPS for at least 15% off any annual membership.

Cheat Sheet

• Why your “I'll learn AI later” plan is already backfiring [1:12]
• Top 3 Productivity Resources [2:43]
• The surprisingly useful thing you can do with ChatGPT in your car [4:30]
• What Google's 1998 launch tells us about AI's tipping point right now [5:55]
• It's not AI taking your job — it's the person who knows AI [9:13]
• The college students who can't get hired until they build something [11:35]
• The two everyday skills that predict how fast you'll pick up AI [24:15]
• Why using the tools daily matters more than watching tutorials [26:08]
• Think of AI like a new hire — it's bad at first and that's the point [26:56]
• Stop copy-pasting AI drafts into Gmail — here's what to do instead [29:40]
• How Thanh trained Claude to write emails that sound like him [30:52]
• The curiosity trick that unlocks capabilities you didn't know AI had [32:13]
• Why asking “can AI do this?” beats Googling how to do it yourself [36:57]
• Wrote off AI after one bad experience? Time to try again [38:46]
• The side quests that turned into a real client project a year later [43:34]
• You don't have to build apps — small daily wins count too [47:12]
• From 30-minute manual updates to a fully automated portfolio tracker [49:30]
• Your Mac Mini might become your always-on AI agent server [53:36]

Links

• Shopify
• SelectQuote.com
• Notion Custom Agents.
• Masterclass
• 25X Productivity Coaching
• Claude Cowork
• Co-Intelligence: Living and Working with AI by Ethan Mollick
• ChatGPT CarPlay App
• Google Maps Ask Maps Feature
• Lovable
• Productivity Academy
• The Productivity Show on Youtube
• The Productivity Show Youtube Clips
• Become a member of TPS+
• Take the Productivity Quiz

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

The pace is fast. Something meaningful happens almost every week. If you are using AI seriously, falling behind by even a month means you are missing context that actually matters for your work.

Most people respond by reading more. More newsletters. More articles are bookmarked to read later, which slowly pile up into a reading queue that never gets shorter.

I did that for a while. It did not work. I was always behind, and every piece of new information just reminded me of five other things I had not gotten to yet.

I do something different now.

The Three-Step Workflow

When I want to stay on top of AI news and updates, I do this:

1. Download whatever I want to absorb: AI newsletters, article roundups, product update notes, PDFs, whatever has accumulated
2. Drop everything into NotebookLM
3. Generate a podcast from the source material, then listen during my workout

That is the whole thing.

By the time I am done at the gym, I have covered everything I needed to. Not skimmed. Actually absorbed. And I have not spent a single extra minute sitting at my desk to do it.

Why NotebookLM Works for This

NotebookLM podcast generation is genuinely good, and I think it is underused for this specific purpose.

You give it source material and it generates a conversational audio discussion based on that content. Not text-to-speech reading. An actual back-and-forth between two voices that discusses, explains, and contextualizes what is in the material.

The quality surprised me when I first tried it. It handles nuance reasonably well. It connects ideas across your different sources. And because it is conversational, it is much easier to follow when you are doing something physical than a dense written summary would be.

I have found it more useful than most actual podcasts on AI topics, because it pulls from exactly what I want to know about, not whatever a host decided was interesting that week.

The Real Shift: Learning Without Separate Time

The deeper reason this workflow matters is not the tool. It is the mindset behind it.

Most people treat learning as a separate activity. Something that requires sitting down, opening a browser, carving out focused time. And that framing is fine for some things.

But staying current is different. That is mostly about exposure and absorption, not deep analysis. And you can do exposure and absorption during time that is already allocated to something else.

I work out anyway. The workout was going to happen. Adding the AI news podcast to it costs me nothing.

This is the One Tweak a Week principle applied to how you consume information. You are not adding a new habit. You are layering something useful onto a habit that already exists. The friction is almost zero because the anchor activity is already happening.

What to Put In It

For source material, I typically use AI newsletter roundups, update notes from tools I use regularly, transcripts from AI talks I have not had time to watch, and occasional long-form articles I want to understand but have not read.

A handful of good sources generates enough material for a 20 to 30 minute podcast. That is about one gym session.

You can also get specific. If you are trying to understand one particular topic, feed it targeted material and generate a focused episode on that.

Setting It Up

NotebookLM is free via Google. Create a notebook, upload your sources, and use the Audio Overview feature to generate the podcast. Takes maybe ten minutes to set up the first time.

From there, the process repeats. When new updates come in, drop them in, generate, listen.

One setup. Ongoing payoff.

The AI space is not slowing down. But saying you do not have time to keep up is increasingly a choice, not a constraint.

Looking for a structured approach to building AI into your workflow one step at a time? The 4-Day AI Sprint is built exactly for that.

Mo wanted to meet about his Latte coffee party. Rusty, my CPA, had a call with Evan at the same time. Two conversations I cared about, same slot, no way to split myself.

I sent my AI notetaker to Rusty call. I physically attended Mo. By the time both calls ended, two complete summaries were in my inbox. Structured summaries of what was discussed, what was decided, what actions were agreed to.

Why This Changes How You Think About Meetings

Most people treat AI notetakers as a convenience. They miss what is actually happening. Your presence has always been the bottleneck to staying informed. The notetaker breaks that constraint. Meetings and calls are assets, not just interactions. The transcript is the lasting artifact.

The Accuracy Problem With Human Debriefs

Transcripts are more accurate than asking someone else what happened. If I asked Evan how the Rusty call went, I would get his version, his memory, his filter. The transcript has everything. Word for word. I thought we said X is a lot harder to float when the other person has a verbatim record.

Building on Top of the Transcript

I have set up a workflow where if someone commits to sending me something on a call and it does not arrive within a couple of days, Lindy drafts a follow-up for me. No manual tracking. The transcript is the source of truth. The agent acts on it. Without the transcript, none of it works. With the transcript, you can automate a lot of what used to require manual attention after a call.

Start With the Recording

If you are not recording your calls, start there. Tools like Lindy, Otter.ai, and others can join a call and transcribe it automatically. Setup is usually five minutes. Once it runs, it runs. Your presence is still valuable. But it does not have to be the only way you get information from a conversation that happened without you.

Want to build a full meeting intelligence system? The 4-Day AI Sprint walks through how to set this up.

We were deep in it — someone had just shared their setup, and a few people were riffing on how they’d handle a specific problem: searching through years of email, calendar data, and text messages to build a prospect list from scratch.

And I said something that caught people off guard.

“I would actually not use Lindy for most of this.”

The room got quiet. Which made sense — we were in a Lindy session.

But here’s why I said it, and why I think this matters for how you think about AI tools in general.

The Wrong Way to Think About Automation Tools

Most people pick a tool they like and then try to make everything fit that tool.

I’ve watched this play out over the years with Zapier. With n8n. With Python scripts. With no-code platforms. Someone discovers a tool that works really well for one thing, then they spend the next six months forcing every problem through that same mold.

It’s understandable. Learning a new platform takes time. It’s easier to get better at one thing than to maintain expertise across several.

But the cost is real. When you use the wrong tool for a job, you either can’t build it at all, or you build something fragile that breaks when something slightly unexpected happens.

The fix isn’t finding the perfect universal tool. It’s building a mental model for routing work to the right tool.

The Framework: Deterministic vs. Exploratory

Here’s the question I ask before building any automation:

Is this task deterministic or exploratory?

Deterministic tasks have predictable inputs and predictable steps. The same trigger happens, the same process should follow. An email comes in, you summarize it and log it. A form gets submitted, you notify three people. A meeting ends, you generate a follow-up. These tasks repeat on a cadence or on a trigger. You can define the logic once and walk away.

This is where Lindy shines. It’s built for rinse-wash-repeat workflows. Set it up, let it run, don’t think about it again.

Exploratory tasks are different. You don’t know exactly what you’ll find when you start. You’re searching through messy, unstructured data. You’re making judgment calls based on what turns up. You need a tool that can reason in real time, adjust its approach, and handle things it wasn’t explicitly programmed for.

Searching through years of emails and texts to build a prospect database? Every step depends on what the previous step found. That’s exploratory. And Claude Code — which can write and run code, iterate based on results, and handle ambiguity — is a much better fit.

A Real Example of Getting This Right

When I was working on connecting different data sources a while back, I spent about two weeks trying to make a Python script do it reliably. It worked in testing. It fell apart in production because inputs weren’t as clean as I expected.

Then I moved the recurring part of that workflow to Lindy. Cost comparison wasn’t close — and the maintenance overhead basically disappeared. Lindy just handled it every time without me touching it.

But I kept Python and Claude Code around for the exploratory jobs. The one-off research tasks, the data cleanup projects, the things where I needed to investigate first and build second.

Both tools are in regular rotation. They handle different jobs.

The RATs Diagnostic

Before building any automation, I run what I call a RATs check. Is this task:

• Redundant — something you do repeatedly on a schedule or trigger?
• Annoying — low-value work that pulls your attention?
• Time-sucking — taking longer than the output justifies?

If yes to all three: it’s an automation candidate.

Then the question becomes which tool. And the deterministic/exploratory framework tells you where to start.

Redundant + predictable = Lindy. Exploratory + one-off = Claude Code. Something in between — a task that repeats but also needs real-time reasoning — might be a hybrid where Lindy handles the trigger and scheduling, and Claude Code does the thinking.

The Broader Point

I keep saying this in workshops: the more tools you know, the more leverage you create.

Not because you use all of them at once. But because you can route work to whichever one fits best.

A carpenter doesn’t use a hammer for everything just because they’re really good with a hammer. They know what a chisel does, what a router does, what a hand saw does. The expertise isn’t in one tool — it’s in knowing which tool the job calls for.

Same logic applies here. Lindy is excellent. Claude Code is excellent. They’re excellent at different things.

The highest-leverage thing you can do right now is build that mental model. Figure out the shape of each tool’s strengths. Then, when a new problem shows up, you’ll know within five seconds which direction to go.

That routing instinct is worth more than being really, really good at any single platform.

Want help figuring out which tasks in your workflow are worth automating first? The 4-Day AI Sprint gives you a systematic way to identify your highest-leverage opportunities.

We all dream of a life where we're thriving, not just surviving. A career that excites us, personal relationships that nourish us, and a constant sense of forward momentum. But how do we get there? It isn't always about working harder or chasing the next big opportunity. Sometimes, the most powerful step you can take is to look inward and invest in yourself.

Think about it. We plan our financial investments carefully, knowing that dollars put in today can yield significant returns tomorrow. But what about the most valuable asset you own: you? Like a well-tended garden, your skills, knowledge, and relationships need regular care. Neglect them, and you risk being overtaken by a rapidly changing world.

My co-host, Brooks Duncan, and I have seen this time and again. We witness professionals who seem to glide from opportunity to opportunity because they have made the commitment to invest in themselves. Whether it's developing new skills or nurturing personal connections, the simple act of investing in yourself is often the secret sauce to success.

Sharpen Your Saw: The Power of Skill Development

When we talk about personal growth, skill development is key. The more you build your skills, the more valuable you become in every aspect of life. At Asian Efficiency, we live by our core value called “Glow Green,” which encourages us to share what we're reading, listening to, and learning every week.

Brooks once shared how attending a digital marketing conference sparked a breakthrough for him. He described it as having “fireworks going off in his mind.” That one experience accelerated his progress and opened up new opportunities. Whether you choose to go wide—acquiring a broad range of skills—or go deep—mastering one area until you're in the top tier—intentional learning is a game changer.

Remember this: combining two top-tier skills can make you a unicorn in your field. For instance, many are skilled in bodybuilding, but those who add video editing and on-camera presence become truly unique. In my journey, my deep knowledge of productivity combined with writing helped kickstart Asian Efficiency. It's about creating leverage that sets you apart.

Cultivate Your Connections: Building a Strong Network

Beyond developing skills, the relationships you build are an equally important investment. I once read a book called “30 Lessons for Living” that reminded me that our happiness largely comes from our connections with others. I realized I had many shallow relationships—knowing someone's name but not really understanding their story.

That realization drove me to host dinner parties and events. Not only did this make me happier, but it also opened up opportunities I never expected. I've been approached for job offers, invited into exclusive deals, and even met influential figures like Michael Dell and Gary Vaynerchuk. Networking is not just about career advancement; it's about enriching your life with mutually beneficial relationships built on trust.

Here are a few ways to start building your network:

• Consistently share value on platforms like LinkedIn.
• Organize simple events such as monthly brunches or dinners.
• Attend meetups, conferences, or industry groups where like-minded people gather.
• Be relentlessly helpful. The more you give, the more you receive.

Accelerate Your Growth: The Power of Coaching

The third way to invest in yourself is by getting coaching. Most of us know what we should do deep down, but we often struggle with follow-through. A good coach fills that gap by providing tailor-made advice, holding you accountable, and helping you identify your blind spots.

Take my own experience with fitness. I don't particularly love working out, but having a personal trainer makes all the difference. Similarly, I once worked with a copywriting coach who helped me refine my writing faster than self-study ever could. Top performers like LeBron James rely on coaches for this very reason—they help you update the “software” in your head and accelerate your progress.

At Asian Efficiency, we're excited to offer one-on-one support through our 25X Productivity Coaching program. Whether you've taken our online course or attended a workshop, personalized coaching can help you master your time, energy, and attention, and ultimately, achieve the results you're after.

At a Lindy community session I run monthly, someone asked a question I hear pretty often.

They were setting up an AI agent with a detailed context profile — basically a document that tells the agent everything about them, their preferences, their working style. Around 20 pages. And they wanted to know: wouldn't loading all that into an agent cost too many tokens?

Fair question. Cost is real. But it's the wrong thing to be optimizing for right now.

Here's the analogy I used.

Remember the T-Mobile “Top Five” Plan?

For those who were texting in the mid-2000s, you had to pick five people. Five contacts you could text for free. Everyone else costs money. You thought carefully about that list. Family. Best friend. Maybe someone you were dating.

And you rationed your messages to everyone who didn't make the cut.

Now you text anyone without thinking. iMessage, Android, doesn't matter. You don't know what texts cost per message. You probably never think about it at all.

That's where AI compute is heading.

Right now, we're paying early-adopter prices. Just like we paid per-text in 2005. The compute cost is real, but it's a temporary condition, not a permanent constraint. Every few months, the major model providers cut prices. Processing that would have cost $9 per query in 2024 costs $0.07 now. That trend is not slowing down.

In a few years, most people won't think about token costs any more than they think about texting costs.

What My Context Profile Looks Like

My personal context profile — the document I load into my most important agents — is 20 pages. It covers my working style, my communication preferences, my priorities, key relationships, how I like to make decisions, and dozens of other details.

I've seen the CEO of Lindy's version. It's even longer.

Neither of us has noticed meaningful performance degradation from loading it. Neither of us stresses about the token cost.

Does a longer profile cost more? Yes. But “a little more” on something that already costs very little is still very little.

The context profile is what makes an agent actually feel like it knows you. Without it, you're prompting a stranger every time. With it, you're working with something that has real context about who you are and what matters to you.

The ROI on that is hard to argue with.

When Cost Actually Matters

I don't want to pretend tokens are free. There's a real scenario where cost optimization matters a lot.

When you're running thousands of operations at scale, a small per-operation cost compounds fast. If you're enriching a list of 5,000 leads and each record costs $2 to process, that's $10,000. If you can get the same quality at $0.07 per record with better architecture, you just saved almost $10,000.

That's worth optimizing.

The rule I use: when you're doing something thousands of times, optimize hard for cost. When you're doing something daily or handling your own workflows, don't bother. The marginal cost of your time thinking about it exceeds the savings.

What to Focus On Instead

The question I try to redirect people toward is this: what's the highest-leverage thing you could automate next?

Not “how do I spend fewer tokens on this?” but “what task is eating my time every week that an agent could handle—so I can reclaim my calendar?

Those are very different questions, and they lead to very different places.

The first one keeps you in optimization mode, squeezing drops out of a bucket that's getting cheaper every month anyway.

The second one keeps you in growth mode, finding new leverage every time.

If you're at the stage where you're worried about token costs on your personal agents, you're probably not yet at the stage where it actually matters. Focus on building first. Optimize when scale demands it.

We're still early. The text plan era of AI is temporary. Use that window to learn fast and build fast.

The costs will sort themselves out.

Want a practical framework for knowing what to automate first? The 4-Day AI Sprint walks you through it — four days of high-leverage AI skills.

Before I finished, one of the investors had his phone out, pulling up Pinterest, and saving coffee shops he liked.

“I've been trying to figure out what aesthetic I want for our new office space for three months,” he said. “I think I just did it in ten minutes.”

This is one of those workflows that seems almost too simple. Three steps. Costs nothing. And it does something that traditionally costs $10,000 and takes weeks.

The Problem It Solves

When you're building something visual — a brand, a space, a product line — you need to be able to describe what you want. Not just “I like this vibe” but the actual vocabulary: the design style, the lighting, the color relationships, the era, the influences.

Without that vocabulary, you're stuck. AI image generators give you generic outputs when your prompts are vague. Designers charge by the hour to extract your vision through lengthy discovery sessions. And you end up going back and forth, trying to articulate something you feel but can't quite name.

The bottleneck isn't taste. Most people have perfectly good taste. The bottleneck is translation — turning what you see and feel into words precise enough to brief a designer or prompt an AI.

The Three Steps

Step 1: Build a Pinterest board of your aesthetic.

Go to Pinterest and search for the type of space, brand, or product that feels like what you're going for. If it's a coffee shop, search “coffee shop interior” and save the ones that feel right. If it's a clothing brand, search competitors and inspirations. If it's a home or office space, look up interior designers or hotels you admire.

Save 20 to 30 images that represent the target aesthetic. Don't overthink it. If it resonates visually, save it.

Step 2: Upload to ChatGPT and ask it to name the pattern.

Create a new chat in ChatGPT. Upload the images (or the board if you can export it). Then ask something like: “What do all of these have in common? What's the design language, the aesthetic style, the visual elements? Give me the exact terms I'd use to describe this to an AI image generator or a professional designer.”

ChatGPT will tell you. It might say something like “Japandi-influenced with wabi-sabi elements, warm neutrals, natural materials, diffused northern light, and a mix of organic and geometric forms.” Or it'll point to a specific architectural era, a regional influence, a photographic style.

These are the words you needed. And finding them used to require either years of design education or paying someone who had it.

Step 3: Use the vocabulary.

Now you have precise language. Drop those terms into your AI image prompts and watch the quality of outputs jump. Use them in your brief to a designer — they'll spend less time guessing and more time executing. Add them to your brand guide so anyone on your team can create on-brand visuals.

The investor who pulled out his phone at my workshop walked away with a design brief for his office in the time it took most people to answer their email.

Why This Works

This is essentially reverse-engineering expertise. A good designer's value is partly their taste, but a huge part of it is vocabulary. The ability to identify and name what makes something look the way it does. That translation layer is expensive to buy and slow to develop on your own.

ChatGPT has seen enough design photography, architecture, and visual content to recognize patterns. It can't generate taste from scratch, but it's very good at pattern recognition. At naming what you're already looking at. At bridging the gap between your visual intuition and the precise language needed to brief an AI or a human.

I think of this as building your Camera Language for Images — the vocabulary needed to direct an image AI like a director briefing a cinematographer. Once you have it, your prompts stop being guesses.

One Good Use Case You Might Not Have Thought Of

This workflow isn't just for brand design.

If you're pre-selling a product that doesn't exist yet — using AI-generated photorealistic images to validate demand before manufacturing — you need your images to look extremely specific and compelling. Vague prompts give you vague renders.

Run this workflow first. Get your design vocabulary. Then use it in your product rendering prompts. The difference in output quality is real.

Start Here

This week, pick something you've been meaning to nail down visually — a brand, a space, a product, a presentation aesthetic.

Go to Pinterest. Spend 15 minutes saving what resonates. Upload to ChatGPT. Ask it to name the pattern.

You'll have your design brief in half an hour. For free.

Want to go deeper on AI tools and workflows that actually save money? Check out the 4-Day AI Sprint — four days of practical AI skills you can use immediately.

Midway through, I showed them a prompt I'd been using for interior design renders: “Japondi style, diffused lighting, modern contemporary twist, corner angle, 50mm lens.”

Dead silence.

One person finally said, “I don't know what half of those words mean.”

That was the point.

When the AI Got Good, You Became the Bottleneck

Image generation models have gotten extraordinary. If you haven't tried Gemini's image model or Midjourney recently, give it ten minutes. The quality is there.

The problem isn't execution anymore. The problem is description.

When you type “make this space look cool,” the AI will generate something. It might even look fine. But “something” and “exactly what you had in mind” are very different things.

And the gap between those two outcomes lives entirely in your vocabulary.

This is what I call Camera Language for Images. It's treating an image prompt the way a director treats a shot list. Focal length. Lighting style. Aesthetic reference. Camera angle. Once you start thinking in those terms, your prompts stop being guesses and start being directions.

What I Saw at the Paddle Club

A friend of mine spent about 30 minutes with AI recently to build a complete brand identity for his paddle club. Logo system, color palette, design language, the whole thing. He showed me the outputs. They were genuinely good.

Traditional agency work would have been $25,000 to $40,000 and taken months.

But here's the thing I keep coming back to: he's been studying design for years. He has taste. He knows what good looks like and, more importantly, he knows how to describe it. The AI moved fast on what he already knew how to ask for.

That's the part most people try to skip.

You don't get great AI outputs by prompting harder. You get them by knowing what you're looking for well enough to describe it precisely.

The Vocabulary Gap Is a Skill Gap

Most people treat image prompting as a copy-and-paste game. Find a prompt on Reddit. Swap out the subject. Hope for the best.

And sometimes it works. But it breaks down the moment you want something specific, because you're borrowing someone else's vocabulary for someone else's vision.

The faster path is building your own vocabulary.

I built a small daily practice loop with Lindy. It sends me a photo each morning and asks me to identify the design style. Japondi. Wabi-sabi. Scandinavian. Bauhaus. Mid-century modern. After a few months of that, those words stopped being foreign and started showing up naturally in my prompts.

Then something shifted. The gap between what I pictured and what the model generated got very small.

The Three Areas Worth Learning

You don't have to go deep on all of this at once. Here's where to start:

Design aesthetics. Pick two or three named styles and learn what they look like. Japondi (Japanese-Scandinavian fusion) is a good one. So is Wabi-sabi, Bauhaus, and organic modern. Pinterest is useful here. Search a style name and spend 15 minutes looking at the images. That's enough to start using the name in prompts.

Photography language. Focal length changes the feel of a shot dramatically. A wide-angle (24mm) makes a space feel expansive. A 50mm feels natural and close. An 85mm compresses and flatters. Lighting terms like diffused, hard light, golden hour, and softbox all change the mood. You don't need to be a photographer. You need to know the words.

Camera angle and framing. Corner angle, low angle, eye-level, bird's eye. These are two-word phrases that completely change what the AI generates. Learn four of them and use them.

One Week to a Better Prompt

Here's the smallest possible place to start.

Pick one design style this week. Look it up on Pinterest. Save five images that represent it well. Then write one image prompt that uses the style name, one lighting term, and one camera angle.

See what comes back.

It won't be perfect the first time. But it'll be more specific than “make it look cool.” And specific is where the good stuff lives.

The image AI isn't the bottleneck. Your vocabulary is. And vocabulary is something you can build, a little bit at a time.

Want to go deeper on AI tools that actually move the needle? Check out the 4-Day AI Sprint to get up to speed fast.

I was on a call with Dieter a couple of months ago. He's deep into AI agents… runs a bunch of them for his architecture visualization business. But he had a problem I recognized immediately.

“Every time my priorities change, I have to go into each agent and rewrite the prompt. It takes forever.”

I pulled up my screen and showed him something. A Google Doc. Nothing fancy. About 3 pages. I call it the shared brain.

The Problem With Multiple Agents

Here's what happens when you start building out a real agent system. You get an email agent. A content agent. A meeting prep agent. A research agent. Maybe a CRM updater. Each one has its own prompt, its own instructions, its own idea of what matters.

And when your priorities shift… which happens every week if you're running a business… those agents don't know. Your email agent is still prioritizing responses to old threads. Your content agent is writing about topics you've moved past. Your meeting prep agent doesn't know which projects are active and which are on hold.

You end up spending more time managing the agents than the agents save you. That's backwards.

What the Shared Brain Actually Is

It's a single Google Doc that every agent reads before it does anything. That's it.

The doc contains:

• My current top 3 priorities (updated weekly)
• Active projects and their status
• KPIs each agent should optimize for
• Which types of tasks are urgent right now
• Any temporary rules or exceptions

So when I update the file to say “AI consulting implementation is the main priority this quarter,” every agent shifts behavior. Email drafts prioritize conversations with consulting clients. Content angles toward AI implementation topics. Meeting prep pulls context for those specific calls. CRM tagging flags consulting-related contacts.

One update. All agents aligned. Instantly.

This is what I think of as the agent-as-teammate model. The long-term goal isn't a bunch of isolated bots following rigid scripts. It's a team of agents that are aware of your priorities, your history, and your context. The shared brain is how they stay coordinated.

Why It's Embarrassingly Simple

The thing that surprises everyone when I show this: it's not a database. It's not some fancy multi-agent coordination framework. It's not RAG or vector embeddings or any of that.

It's a Google Doc.

But it works for two reasons. First, every modern AI tool can read a Google Doc. Lindy connects to it natively. You can paste it into Claude or ChatGPT. It's the most portable format there is.

Second, one source of truth beats distributed truth every time. When agents all read from the same file, they can't get out of sync. When you update it, you update it once. Done.

I learned this from managing real teams, honestly. The best-run companies I've seen all have something like this… a one-page strategic priority doc that everyone references. The shared brain is just the AI version of that.

A Real Example

Here's how it played out last month. I shifted my focus from workshop planning to consulting implementation. Took me about 5 minutes to update the shared brain.

Within the next cycle, my email agent started drafting more detailed responses to consulting prospects. My content agent shifted toward implementation-focused topics. My meeting prep agent started pulling relevant case studies before consulting calls.

I didn't touch a single agent prompt. Just updated the doc.

Dieter's reaction on the call: “Wait, that's it? I've been spending 30 minutes rewriting prompts every time something changes.”

How to Build Your Own

If you're running 2+ agents and they feel disconnected, here's how to start:

1. Open a new Google Doc. Title it something like “Agent Priorities” or “Shared Brain.”
2. Write down your top 3 priorities for this week or quarter. Be specific. Not “grow the business” but “close 3 consulting deals by end of March.”
3. List your active projects with a one-line status for each.
4. Add any rules or preferences your agents should follow. “Always draft, never send.” “Prioritize Austin-based contacts.” Whatever matters right now.
5. Connect this doc to every agent you run. In Lindy, add it to the Knowledge Base. In other tools, paste it into the system prompt or reference it as a file.
6. Update it weekly. Takes 5 minutes during your weekly review.

That's the whole system. No code. No database. Just a document that keeps everyone… human and AI… on the same page.

The Bigger Picture

The shared brain is actually the minimum viable version of something bigger. As your agent system grows, you can layer in per-agent role definitions, person-specific dossiers, and decision logs. But the shared brain is always the foundation.

Start with the text file. Get your agents reading from one source. Everything else builds on top of that.

Was on a call with Brooks a few months ago. He's my podcast co-host and we were talking about AI personalization. He'd been testing agents for email drafting and kept getting outputs that sounded… fine. Correct tone. Decent structure. But generic. Could've been anyone.

I showed him something. Pulled up a Google Doc. 35 pages long. And told him every single agent I run reads this document before it does anything.

His first question: “How did you build that?”

What's In the Document

It's everything about me. My name, where I live, my time zone. My bio. My current revenue streams and what I'm focusing on this quarter. My core values and principles. How I make decisions. What my brand voice sounds like. Communication preferences. Even the businesses I'm running and what stage each one is at.

It's not a resume. It's more like a really thorough operating manual for “how to be Thanh.” The kind of document that, if you read it carefully, you'd know exactly how I'd respond to any given situation.

I think of these as context files… reusable text documents that encode who you are in a way AI can use. They're assets. You build them once, load them into every workflow, and the output quality goes up across the board.

How I Actually Built It

Here's the part that surprises people. I didn't sit down and write 35 pages from scratch. That would take forever.

I do a weekly review every Sunday. Have been doing it for years. Each review covers what went well, what didn't, what I learned, what I'm focusing on next week. Pretty standard productivity stuff.

I took all 44 weekly reviews from 2025 and pasted them into ChatGPT. Literally just dumped them in and said something like: “Combine everything here into a comprehensive profile of who I am. Include my values, priorities, communication style, business focus, and decision-making patterns.”

35 pages came out the other side.

Then I cleaned it up. Removed things that were outdated. Added a few sections about my brand voice and writing style. But the bulk of it came straight from a year's worth of self-reflection that I'd already done.

What Changes When Agents Have This

The difference is immediate. Without the context profile, AI drafts emails that are technically correct but could be from anyone. With it, the drafts sound like me. My email agent knows I write short, casual messages. It knows I use ellipsis instead of em dashes. It knows I prefer direct language over corporate speak.

My content agent knows which topics I care about right now. It knows my stories. It knows the analogies I tend to reach for.

My meeting prep agent knows my priorities without me re-explaining them. It knows which projects are active, which ones are on hold, and what metrics I track.

All because every agent reads the same 35-page document before it starts working. Centralized context. One source of truth for all of them.

Why Most AI Feels Generic

Here's the insight that clicked for Brooks on that call. He said: “So it's not about the prompt. It's about the input.”

That's exactly it.

Most people spend all their energy crafting the perfect prompt. And the prompt matters… but it's maybe 10% of the output quality. The context you give the AI is the other 90%.

Generic prompt plus no context equals generic output. Decent prompt plus 35 pages of context about who you are, how you think, and what you care about… that equals output that actually sounds like you.

How to Start Building Yours

You don't need 44 weekly reviews. Start with what you have.

If you already do weekly reviews: Upload the last 3-6 months into ChatGPT or Claude. Ask it to synthesize a profile of who you are, what you prioritize, and how you communicate. You'll get a solid starting point.

If you don't do reviews: Start with a brain dump. Spend 30 minutes answering these questions:

• What do you do for work? What are your current projects?
• What are your top 3 priorities this quarter?
• How do you prefer to communicate? Short and direct or detailed?
• What values guide your decisions?
• What does your typical week look like?

Save that as a document. Even 2-3 pages is better than nothing. You can grow it over time.

The key principle: Keep it in a readable format. Plain text or Google Doc. Something any AI tool can ingest. Don't lock your context inside one platform.

The Compounding Effect

Here's what's kind of wild. The document gets better over time. As I do more weekly reviews, I update it. As agents learn new things about my preferences, some of them can actually update the profile themselves.

Six months from now, the document will be even more accurate. A year from now it'll know me better than most people do.

That's the real play. Build this once. Keep feeding it. And every AI tool you use from now on starts with a massive head start on understanding who you are.

Your weekly reviews aren't just for reflection anymore. They're training data for your AI.

Asked a room of investors a question back in January. “How many of you know how to drive a car?”

Every hand went up.

“How many of you actually know how the electrical system inside your car works?”

Maybe two hands. Out of sixteen people.

And that's the analogy I keep coming back to when I talk to business owners about AI. We all drive cars. Every day. We commute, we run errands, we take road trips. We get enormous value from driving. And almost none of us know how the engine actually works. We don't need to. We just need to know how to operate the thing.

AI is the same.

The “I'm Not Technical Enough” Problem

There's a misconception floating around right now that you need to be technical to use AI well. That you need to understand how large language models work. That you need to know what tokens and embeddings and fine-tuning mean. That you need to be able to code.

You really don't.

I say this as someone who uses AI every single day for high-value work. Drafting content. Researching prospects. Automating follow-ups. Building workflows. And honestly… I don't know on the technical level how these LLMs work. Not really. I know enough to use them well. That's different from understanding the engineering.

It's the same way I know enough about my car to fill the gas tank and check the tire pressure. I don't need to rebuild the transmission.

The Gap Nobody Talks About

Here's something that surprised me. If you spend time on AI Twitter or YouTube, you'd think everyone already gets this. You'd think every business owner is building agents and automating workflows.

They're not. Not even close.

I had a conversation with a friend who runs a marketing community. He pointed out the massive gap between what the tech crowd knows and what actual business owners know. Smart, successful people who are busy running companies… and they barely know how to write a basic prompt. They don't know what a custom GPT is. They've never heard of Claude or Lindy.

When you're stuck in the AI bubble, you assume everyone knows these things. But the reality is there's this huge middle ground of capable people who just need someone to show them the basics.

What I See at Workshops

Taught salon owners at Sirius Business last month. Three sessions. All standing room only. These aren't developers or tech workers. They're people trying to fill chairs, post better photos on Instagram, and manage appointments.

Most had never opened ChatGPT before walking in.

By the end of the session, they were using it. Not because I taught them how neural networks function. But because I showed them how to do one specific thing they already needed to do… just faster. Sort product photos. Write captions. Draft responses to reviews.

The barrier was never technical ability. It was fear. People think they need to be experts before they can start. They don't. They just need to try one thing.

The Three Levels (And Why Level One Is Enough to Start)

I teach AI in three levels. Level one is AI-assisted… you use tools like ChatGPT to help with tasks you already do. Level two is workflows… you start connecting tools so things happen automatically. Level three is agents… AI systems that work semi-independently.

Most people haven't even started Level One. And that's fine. Level One alone can save you hours every week.

You don't need to jump to building agents. You need to learn to drive first.

Start Here

Pick one thing you do every week that takes too long. Maybe it's writing emails. Maybe it's summarizing meeting notes. Maybe it's researching prospects.

Open ChatGPT, Claude, or Gemini. Describe what you need. See what comes back.

That's it. That's how everyone starts.

You'll get better fast. Not because you'll suddenly understand machine learning. But because you'll learn what to ask for and how to ask for it. Just like driving… you get better by doing it, not by studying engine diagrams.

The car doesn't care if you know how it works. It just drives.

Was showing a client my content system a few months ago. He runs an executive training business and creates a lot of thought leadership content. Like most people, he'd been using ChatGPT to draft posts. And like most people, the results were… fine. Technically correct. But generic. Could've been written by any consultant in his space.

I showed him something different. I pulled up a LinkedIn post my system had drafted about leadership. It referenced a specific conversation I'd had with an investor in December, where I explained how one text file coordinates 14 different AI agents. Real conversation. Real detail. Not something any LLM would invent.

His first question: “How does it know that story?”

The answer is something I call the story bank.

What a Story Bank Actually Is

Every meeting I'm on, a bot captures the full transcript. That part's pretty standard now. Most people stop there… they get the transcript, maybe a summary, and it sits in a folder somewhere.

I take it further. An agent processes each transcript and extracts every story I told. Not the agenda items or action points. The stories. The anecdotes. The real examples from my actual life and business.

Those stories go into a Supabase database. Each one gets tagged with keywords and embedded as a vector so I can search them semantically later.

This is what I mean by transcript-first thinking. The transcript isn't just a record of what happened. It's raw material. Every conversation you have is full of stories you've already told naturally… you just need a system to capture and reuse them.

Why Vector Search Changes Everything

I started with Google Sheets. Had a column for the story, a column for keywords, and I'd manually search through them. It worked for maybe 50 stories. After that it got slow and messy.

The real unlock was switching to Supabase with vector search. Here's the difference.

With keyword matching, if I search “leadership,” I only find stories where the word “leadership” appears. But some of my best leadership stories never use that word. I might talk about a time I let a team member make a decision I disagreed with. That's a leadership story… but keyword search would never find it.

Vector search understands meaning, not just words. Search “leadership” and it finds stories where I demonstrated leadership, even if I never said the word. That's the difference between a useful database and a filing cabinet.

How It Plugs Into Content Creation

Now when my content agents write something… say a post about delegation… they search the story bank first. And instead of manufacturing a generic example about “a busy executive who struggled with delegation,” they pull a real story from a call I had in January where I told a client about automating my entire follow-up process.

That's real. That happened. No LLM invented it.

The result is content that sounds like me because it literally comes from things I've said. Not a style guide. Not a tone description. My actual words and experiences, just reorganized around a new topic.

The Input Problem Nobody Talks About

Here's what I think most people get wrong about AI content. They blame the model. “ChatGPT writes generic content.” “Claude doesn't capture my voice.” “The AI sounds robotic.”

The model isn't the bottleneck. The input is.

If you give AI a generic prompt like “write a LinkedIn post about productivity,” you'll get generic content. Every time. Doesn't matter which model you use.

But if you give it your actual stories, your specific examples, your real conversations… it has something to work with. The output can only be as specific as the input.

How to Start Building Your Own

You don't need to build my exact system. Start simpler.

1. Record your meetings. Use any notetaker… Otter, Lindy, Fireflies. The tool doesn't matter as long as you're capturing transcripts.
2. At the end of each week, pull the transcripts and ask ChatGPT: “Extract every story, anecdote, or specific example from these transcripts. Include the context and key details.”
3. Save those in a spreadsheet or doc. Even a simple Google Sheet with a “Story” column and a “Topic” column is enough to start.
4. Before writing any content, search your story bank first. Find a relevant story and build the post around it.

That's it. You can get fancy with vector databases and automation later. The important thing is capturing your stories before they disappear into forgotten meeting recordings.

The Real Takeaway

Every conversation you have is content waiting to happen. The stories you tell naturally in meetings… those are the most authentic pieces of content you'll ever produce. You just need a system to remember them.

Build the story bank. Let your AI use real material instead of making stuff up. Your content will sound like you because it actually is you.

Was on a call with a friend last month. He runs a marketing community, and we were talking about lead magnets. You know the drill… free PDF, checklist, “ultimate guide.” The stuff everyone gives away to build an email list.

He said something I haven't stopped thinking about. “Tools are the new lead magnets.”

Not ebooks. Not templates. Actual AI-powered tools you give away for free.

Why PDFs Stopped Working

Think about the last time someone gave you a free ebook in exchange for your email. Did you read it? Probably not. It's sitting in your downloads folder right now along with 47 others.

PDFs had their moment. But the bar for what feels valuable has gone way up. People don't want to read your 30-page guide. They want something that helps them right now.

A tool does that. You give someone a tool, they use it, and they get a result. In that moment, they trust you more than any PDF could accomplish.

A Real Example

A friend of mine built an AI ad generator using Google AI Studio. Took him about five minutes. Not a weekend. Not a sprint. Five minutes. It's a simple tool… you put in a few inputs about your product and it generates ad copy.

He gives it away for free. People use it, get decent ad copy, and now they're on his list. They already trust him because the tool actually worked. It delivered value before he ever asked for anything.

This is what I call the value funnel approach. Trust-building value comes before the ask. The best pipeline looks like repeated value-add rather than obvious lead capture. And AI tools are the fastest way to deliver that value right now.

The Economics Are Wild

Here's what makes this moment different from two years ago. The cost to run AI tools has dropped to almost nothing.

Models like Gemini Flash cost fractions of a cent per use. So your lead magnet doesn't just attract people… it costs you basically nothing to maintain. Compare that to running paid ads to a landing page with a PDF download.

And building the tool itself? If you can write a decent prompt, you can build a useful AI tool. Google AI Studio, Claude, ChatGPT… all of them let you create simple tools with no code.

What Kind of Tool Should You Build?

The tool doesn't need to be complex. In fact, simpler is better. A prompt wrapper that solves one specific problem for your audience.

Some ideas depending on your niche:

• An SEO title checker that scores your blog post headlines
• A meeting summary tool that extracts action items
• A follow-up email writer that drafts replies based on meeting notes
• A job description analyzer that flags missing elements
• A social media caption generator for a specific industry

Pick the one thing your audience asks you about most. Build a tool that answers that question for them. Give it away.

The New Playbook

The old content marketing playbook was: create content, build trust, make a sale. That still works. But there's a faster path now.

Tool, instant value, trust, sale.

Same concept, just compressed. Because the tool proves you can help before you ever pitch anything. It's the difference between telling someone you're good at something and showing them.

Start This Weekend

If you've got a niche audience and 30 minutes… build a simple AI tool and give it away. Don't overthink it. The first version doesn't need to be perfect.

Start small. Put it in front of 10 people. See if they use it. If they do, you've got something. If they don't, ask why and rebuild.

The window for this is still open. Most businesses haven't figured out that AI tools are lead magnets yet. The ones who do it first in their niche will have a real edge.

Build the tool. Give the value. Let it do the selling.

He runs a marketing community. Smart guy. Built and sold companies. We were talking about whether to build software products using AI tools. Everyone's doing it right now. Vibe coding is a thing. You can ship an app in a weekend.

His take was blunt: “If you go compete with all the people building apps 24/7, your median income drops. Code is basically free now.”

I've been thinking about that line ever since. Because he's right. And I think most people in the AI space have this completely backwards.

Everyone's Building. Nobody's Distributing.

Right now, the default move in the AI world is: learn to build stuff. Ship an app. Launch a tool. Fork an open-source project and slap a UI on it. And the tools make it genuinely easy. You can build almost anything for almost nothing.

But building something nobody uses is easy. It's always been easy. The hard part was never the code. Even before AI, the hard part was getting people to care.

Now that code is basically a commodity… that gap is even wider. The supply of software just exploded. But the supply of trust didn't.

What Actually Wins Right Now

Look at the businesses doing well in the AI era. Not the venture-funded ones burning cash. The profitable ones. What do they have in common?

They have distribution. An audience. A brand. People who already trust them.

I see this in my own business every day. My workshops sell out not because I built the best AI product. They sell out because I spent years showing up. Writing content. Teaching for free. Building relationships in Austin. When I say “this tool works” or “this workflow saves time,” people believe me. Because they've seen me do it.

That trust is the moat. Not the code.

A Real Example

A friend of mine wanted to help a medical office brokerage reach building owners who might want to sell. Dermatologists, dentists, doctors.

We automated the whole thing with AI. Mined owner data into Google Sheets. Used agents to message prospects on LinkedIn. Before every sales call, the system automatically pulled comparable sales data for that specific property. The salesperson just needed two minutes to review the comps and they were ready.

Impressive system. But here's the thing… it only worked because those salespeople already had credibility in that market. People picked up the phone because they recognized the company. The AI amplified what was already there. It didn't create trust from zero.

If you gave that same automation stack to someone with no reputation and no relationships… it would be spam.

The Counterintuitive Move

So here's what I'd tell anyone trying to figure out where to invest their time right now.

If you're choosing between learning to code AI products and learning to build an audience… build the audience first.

Not because code doesn't matter. It does. But because the bottleneck has shifted. Code used to be the hardest thing to do in business. It was expensive, slow, and required specialized talent. That made it the moat.

Now code is cheap and fast. What's expensive and slow is building trust. Getting attention. Creating a brand people actually care about.

What This Means Practically

A few things I'd focus on if I were starting from scratch:

Create content in public. Share what you're learning, building, and testing. Be specific. Show your work. People follow process, not polish.

Build relationships in a niche. Don't try to be famous on the internet. Try to be known in one room. One industry. One city. That's where trust compounds fastest.

Use AI to scale what's already working. Once you have distribution, AI is incredible at amplifying it. Automating follow-ups. Personalizing outreach. Creating content faster. But it needs something real to amplify.

Don't compete on product. Compete on positioning. If ten people can build the same app, the one who wins is the one people trust. That's not a code problem. That's a brand problem.

The Bottom Line

Play the people game. Build trust. Build an audience. Build a reputation in your space.

Then use AI to scale it.

Not the other way around.

Visit www.asianefficiency.com for more productivity tips and tactics.

Get 20% off your first order: dripdrop.com and use promo code tps.

Links:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Dripdrop
• 25X Productivity Coaching

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

Not the cartoonish 3D models you see in most pitch decks. Not the Sims-looking stuff that traditional rendering software spits out. Actual photorealistic images that people in the room couldn't tell apart from real photos.

The architect's reaction was immediate: “We're going to sell way more now. Because when we go to clients, we can show them what their building will actually look like.”

And he was right. That's the whole game for architecture firms. They're asking people to spend millions on something invisible. Blueprints are abstract. Even fancy 3D models still feel fake. But a photorealistic image of exactly what the finished space looks like… that changes the entire sales conversation.

How I Actually Did This

I've been working on a members club project called Arena Hall here in Austin. We needed to show investors and potential members what the space would look like before construction started. Traditional rendering firms charge $25,000-40,000 for this kind of work. And it takes weeks.

I spent about thirty minutes with a blueprint, a mood board, and a couple of reference photos. Fed them into Gemini's image model. Out came images that looked like someone had walked into the finished building and taken photos.

But here's the part most people miss. The quality of those renders didn't come from the tool. It came from how I described the shot.

Camera Language Is the Skill

I spent hundreds of hours learning what I call camera language. It's basically learning photography… except your camera is a text prompt.

Instead of typing “beautiful modern bar” (which gets you generic clip art), you describe the shot like a photographer would:

• 35mm focal length, eye-level shot from the entrance
• Warm natural light coming from the left, ambient glow from pendant fixtures
• Shallow depth of field, foreground elements slightly blurred
• Rich wood tones, brass accents, dark leather seating

That level of specificity is what makes the output look real instead of AI-generated. The model knows what a 35mm lens looks like. It knows how warm light behaves. It knows what shallow depth of field does. You just have to speak its language.

Then I Turned the Images Into Video

Once I had the still renders, I used Gemini's video generation to create dynamic walkthroughs. Slow pan from the entrance to the bar. Zoom into the lounge area. Each shot described with camera movement language just like you'd tell a cinematographer.

The result was a one-minute video walkthrough of a space that doesn't exist yet. It looked like someone filmed it with a professional crew.

That video became a sales tool. It helped close early memberships for Arena Hall because people could actually see what they were buying into.

Why This Matters Right Now

Six months ago, the models weren't good enough for this. The images looked obviously AI-generated. Faces were wrong. Lighting was flat. Proportions were off.

Now they're good enough. And most architecture firms, real estate developers, and interior designers haven't figured this out yet.

The firms that learn camera language first are going to have a real advantage. They can produce in thirty minutes what used to cost $25,000-40,000 and take weeks. They can iterate on designs in real time during client meetings. And they can pre-sell projects that haven't broken ground yet.

Who Should Learn This

If you sell anything before it physically exists, this is for you:

• Architecture firms pitching new builds
• Real estate developers seeking investors
• Interior designers showing clients what their space will look like
• Event producers showing sponsors what the venue will feel like
• Product designers pre-selling before manufacturing

The common thread is the same. You're asking someone to commit money to something they can't see. Photorealistic AI renders solve that problem.

Where to Start

You don't need to spend hundreds of hours like I did. Start with one project:

1. Grab a blueprint or floor plan
2. Collect 5-10 reference photos of the aesthetic you want
3. Open Gemini and describe the shot using camera language (focal length, lighting, angle)
4. Iterate. Change the time of day. Change the lens. Change the angle.

The first few attempts will look off. That's normal. By the tenth attempt, you'll start seeing what's possible.

And once you see it… you'll understand why that architect said what he said.

I didn't prepare a presentation. Didn't make slides. I just opened my laptop and ran a live demo of an AI agent doing something useful for their business.

About ten minutes in, the father stopped me. “Wait… it can do that?”

He immediately started rattling off ideas. Thirty, maybe forty different ways he could use this across his business. Listing automation, client communication, and market analysis. I barely said anything for the next twenty minutes. He was selling himself.

That's when I realized something I should have figured out way earlier.

Nobody Needs More Information About AI

There are thousands of articles explaining what AI can do. Podcasts. YouTube videos. LinkedIn posts (including mine). People aren't short on information.

What they're short on is seeing it work. On something real. Right in front of them.

It's the difference between reading about how good a restaurant is and actually tasting the food. One is interesting. The other changes your behavior.

The Otis Elevator Lesson

A friend of mine brought up Elisha Otis recently. He's the guy who made elevators safe in the 1800s. The problem wasn't the technology. The problem was trust. Nobody wanted to get into a box held up by a rope.

So Otis went to a convention, got on a loaded elevator platform in front of a crowd, and had someone cut the rope. The safety brake caught it. The crowd gasped. And elevators went mainstream.

He didn't explain why elevators were safe. He showed them.

I think about that a lot. Because the AI adoption challenge right now isn't technical. The tools work. The barrier is belief. People don't believe it can do useful things for them specifically… until they see it happen.

What Happens When You Show Instead of Tell

At one of my workshops here in Austin, a chef showed up. Zero technical background. Probably the least tech-savvy person in the room. By the end of the day, she was building custom GPTs that automated parts of her restaurant work.

She didn't need a pitch. She needed to sit down and try it with someone who could answer her questions in real time.

I've seen this pattern over and over. The salon owners who didn't know AI could answer their phone calls. The investor who realized he could automate his deal sourcing. The content creator who watched an agent draft a week's worth of posts in fifteen minutes.

Every single one of them had the same reaction. “I didn't know it could do that.”

How to Actually Get People On Board

If you're trying to get your team, your clients, or your partners excited about AI… here's what I'd do:

1. Pick one problem they already complain about
2. Build or find an AI solution that fixes it (doesn't have to be perfect)
3. Show them live. Don't explain. Demonstrate.

That's it. Skip the slide deck. Skip the company-wide memo about “our AI strategy.” Just show one person one thing that saves them time on something annoying.

I call this approach start small and iterate. Don't try to sell the whole vision at once. Show one win. Then another. Let curiosity do the rest.

Because once someone sees what's possible, you don't have to convince them. They start convincing themselves.

Try This Today

Think about one person on your team who's skeptical about AI. Find one thing they do every week that's repetitive and annoying. Build a quick demo or find a tool that handles it. Then just… show them.

No pitch. No slides. Just “hey, look at this.”

That's the Otis move. And it works every time.

That question comes up at every single workshop. And I've found the best way to answer it is with a simple framework. Three levels.

Level 1: AI Assistant

This is where most people are right now. You open ChatGPT, Claude or Gemini. You type a question. You get an answer. Maybe you ask it to write an email, clean up a paragraph, brainstorm some ideas, or do a SWOT analysis.

If you get really good at this level… good at prompting, good at giving context, good at asking follow-up questions… you can win back five to ten hours a week. That's real. But here's the thing. You're still the one doing the work. Every time. You open the tool. You type the prompt. You wait for the response. You copy it somewhere. You're the engine. AI is just a faster wrench.

Most people stay at Level 1 and think that's all there is. It's not.

Level 2: AI Workflow

This is where you stop doing one-off prompts and start designing processes.

Instead of “let me ask AI to summarize this meeting,” it becomes “every time I have a meeting, AI automatically gets the transcript, extracts the action items, sends me a summary, and updates my task list.”

You set it up once. It runs every time. Without you being in the loop.

The difference is huge. Level 1 saves you time on individual tasks. Level 2 eliminates entire categories of tasks.

A few examples of what Level 2 looks like:

• Every new lead that comes in gets a personalized welcome email drafted automatically
• Every meeting transcript gets turned into action items and follow-ups
• Every invoice gets logged, categorized, and added to a spreadsheet
• Every YouTube video you save gets summarized and the key points sent to your phone

The pattern is always the same: trigger, process, output. Something happens, AI does something about it, and the result lands where you need it.

Level 3: AI Agents

Now we're talking about digital employees.

Level 3 agents don't just follow a set process. They make decisions. They check your email and decide what's urgent. They prep for meetings by researching the person you're about to talk to. They follow up with leads based on where they are in your sales process. They update your CRM after every call.

They work autonomously. Twenty-four seven. And you manage them like you'd manage a team. Review their work. Give feedback. Adjust their instructions.

I run about 8-10 agents like this right now. Email drafting, meeting prep, person research, CRM updates, content creation, weekly scorecards. Most of my morning is just reviewing what they did overnight.

Where People Get Stuck

The jump from Level 1 to Level 2 is where most people stall out. And I get it. Level 1 feels productive. You're getting answers faster than before. It seems like you're using AI well.

But you're still operating one prompt at a time. Still opening the tool manually. Still copying and pasting results. The ceiling on Level 1 is your own time and attention.

Level 2 breaks through that ceiling because the AI works whether you're there or not.

A friend of mine described the gap perfectly. He said there's a massive difference between what people on AI Twitter think everyone knows… and what actual business owners know. Most business owners are still trying to figure out how to write a decent prompt. Not because they're behind. Because they're busy running companies and nobody's given them a clear map.

That's what this framework is. A map.

How to Move Up

If you're at Level 1 and want to get to Level 2, here's where to start:

1. Pick the one repetitive task that eats the most of your week
2. Write down every step of that task in order
3. Figure out which steps AI can handle
4. Connect them using a tool like Lindy, Zapier, or Make
5. Test it. Tweak it. Let it run.

Don't try to automate everything. Just pick one thing. Get it running. Then pick the next thing.

I call this approach one tweak a week. Do one more thing with AI this week than you did last week. Let curiosity compound.

And if Level 3 feels far away right now… that's fine. You don't need agents today. But knowing they exist changes what you aim for. And that's the whole point of having the map.

“I just want to be in the mode where I'm reviewing stuff. And have the agents do most of the work.”

He runs an executive training business. Sixteen years. Strong relationships. Great reputation. But his days looked like this: back-to-back meetings from 9am to 5pm. Then emails piled up. Follow-ups didn't happen. CRM updates fell through the cracks. Action items from calls just… disappeared.

He wasn't bad at his job. He was drowning in the work around his job.

So we built what I call a Digital Chief of Staff.

What a Digital Chief of Staff Actually Looks Like

It's not one agent. It's a team of eight agents running in the background, each with a specific role.

Here's what his system looks like now:

Email Agent — Reads incoming emails and drafts responses in his voice. By the time he opens his inbox in the morning, replies are already waiting. He reviews, tweaks if needed, and hits send.

Meeting Prep Agent — Thirty minutes before every meeting, it generates a briefing doc. Who the person is, their background, what they talked about last time, relevant email threads. One click and he's fully prepared.

Follow-Up Agent — After every call, it reads the meeting transcript and drafts a follow-up email. Action items get extracted and added to his task list automatically.

CRM Agent — Updates contact records based on meeting notes. No more “I'll update the CRM later” that never happens.

Task Dispatcher — Sends him a daily summary of everything that needs his attention. Prioritized. Organized. Ready to review.

The Shift: From Doing to Reviewing

The thing that changed for him wasn't efficiency. It was operating mode.

Before, his morning started with a blank slate. Open inbox. Read everything. Decide what to respond to. Draft responses. Check calendar. Research who he's meeting. Prepare talking points. It was two hours of setup before he could do any actual work.

Now his morning starts with everything already done. Email replies drafted. Meeting briefs ready. Follow-ups sent. CRM updated. His job is to review, tweak, and approve.

That's the shift most people miss about AI agents. It's not about typing better prompts. It's about having the work already started before you show up.

I think of it like the difference between being the person who does the work… and the person who manages the team doing the work. Both are productive. But one operates at a completely different altitude.

What I Run Myself

I have about 8-10 agents running like this for my own work. They cover email pre-drafting, meeting prep, person research, CRM updates, content creation, and a weekly scorecard.

Most of my morning is just reviewing what they did overnight. Some of it's great and I approve it right away. Some needs tweaking. Occasionally I reject something. But the starting point is never blank.

A friend of mine runs a VC firm. He had a human assistant spending twenty hours a week on meeting prep alone. Scanning inboxes, researching people, pulling background info. We replaced that with an AI system that does it in minutes and delivers it thirty minutes before each meeting.

Twenty hours a week. Gone. And honestly… the AI version is more thorough because it doesn't get tired or forget to check LinkedIn.

How to Build Your Own

You don't need eight agents on day one. Start with one.

The highest-impact starting point for most people is meeting prep. It's time-consuming, it's repetitive, and the payoff is immediate because you walk into every meeting actually prepared.

Here's how to start:

1. Pick the agent that would save you the most time this week
2. Define what it needs to know (your calendar, your inbox, your contacts)
3. Give it a clear role and approval rules (always draft, never send without review)
4. Run it for a week. See what it gets right and what needs adjusting.
5. Add the next agent.

The goal isn't to automate everything at once. It's to get to the point where you're reviewing and approving instead of doing and creating from scratch.

That's what a Digital Chief of Staff does. It gives you your mornings back.

The Bottom Line

If you're spending more time doing tasks than deciding on them… that's the gap. The tools exist to fill it right now. And the people who set this up while it's still early are going to have a real edge over the next few years.

Start with one agent. Get comfortable reviewing its work. Then add another. Let it compound.

The mode you want to be in is review mode. And it's closer than you think.

Get 20% off your first order: dripdrop.com and use promo code tps.

Ready to tackle bigger problems? Go to claude.ai/tps.

Visit asianefficiency.com for more productivity tips and tactics.

Cheat Sheet:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Top 3 Productivity Resources [2:27]
• The end-of-day reset that makes tomorrow's you more productive [8:16]
• Why a cordless vacuum changed everything about cleaning habits [11:28]
• The sneaky reason your bookshelf stays neat (hint: it's on camera) [14:00]
• The three levels of clutter — and why the deepest one is hardest to fix [16:13]
• Too many shoes and not enough honesty: when clutter is really about letting go [21:42]
• Why getting a bigger desk (or house) might be the worst thing for your productivity [24:40]
• What downsizing taught Brooks about the junk room trap [26:08]
• The “own less, own better” philosophy that simplifies everything [27:45]
• The two-desk experiment that 99% of Instagram predicted would fail [29:53]
• The thousand-dollar tool in your pocket that replaces a dozen gadgets [31:24]
• One in, one out: the simple rule that keeps clutter from creeping back [32:57]
• Why your app collection might be the biggest source of mental clutter [34:26]

Links:

• DripDrop
• Claude AI
• 25X Productivity Coaching
• Hidden Bar — Free Mac menu bar icon manager
• Bartender — The longtime Mac menu bar manager app Brooks has used for years but is having stability issues with
• Organizing from the Inside Out by Julie Morgenstern — Classic 2004 book on organization systems and the three levels of clutter; Julie has been on the podcast before
• Digital Minimalism by Cal Newport — Book on paring back digital distractions for a more focused life; Cal has also been on the podcast before
• Setapp — Mac app subscription service that includes Bartender
• Dyson Cordless Vacuums — The premium cordless vacuum Thanh's brother uses that inspired the “reduce friction” insight
• Courtney Carver / Project 333 — Minimalist wardrobe challenge: 33 items for 3 months; Courtney has been on the podcast before
• Notion — All-in-one workspace app discussed as a great example of multi-purpose software replacing specialized tools
• Barbara Hemphill
• Productivity Academy
• The Productivity Show on Youtube
• The Productivity Show Youtube Clips
• Become a member of TPS+
• Take the Productivity Quiz

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

I remember the first time I truly traveled alone. It wasn't a grand, planned adventure, but more of a necessity. My friends had to cancel last minute, and I was faced with a choice: ditch the trip or go solo. Hesitantly, I chose the latter. What followed was an experience that reshaped my understanding of myself and the world around me. It was uncomfortable at first, a little awkward even, but it quickly became one of the most liberating things I've ever done.

Many of us dream of exploring new places, but the thought of doing it alone can feel daunting. We imagine loneliness, missed opportunities, or simply not knowing what to do. But what if I told you that traveling solo isn't just about seeing new sights, it's about unlocking a deeper level of self-discovery and confidence? It's about learning to rely on yourself, to navigate the unexpected, and to truly connect with the world on your own terms.

This is a lesson my friend Matt, also known as Nomadic Matt, has mastered over years of traveling to over 100 countries. He's a New York Times bestselling author and a true expert in navigating the world independently. In a recent conversation, he shared some profound insights into why solo travel is a superpower, and how it can transform your life in ways you never imagined.

Why Solo Travel is Your Ultimate Self-Development Tool

When you travel with others, there's a natural tendency to lean on each other. Someone else might handle the navigation, another might pick the restaurants, and you might just go along for the ride. But when you're on your own, all those decisions fall squarely on your shoulders. This isn't a burden; it's an opportunity.

Matt puts it simply: “When you're on your own, you only have yourself to count on.” This forces you to step up. You have to figure out how to get from point A to point B, decide where to eat, and make all the choices. This constant decision-making, even for seemingly small things, builds an incredible amount of confidence. You learn what you like and dislike, and you become more attuned to your own needs and desires. It's a trial-and-error process that refines your understanding of yourself.

Think about it: how often do we go on autopilot in our daily lives? We drive the same routes, follow the same routines, and often let others take the lead. Solo travel shatters that autopilot. Every interaction, every new street, every meal becomes an active choice. This heightened awareness and constant engagement with your surroundings makes you more present and more capable.

Breaking Out of Your Comfort Zone (Even if You're an Introvert)

For many, the biggest hurdle to solo travel is the fear of loneliness or not being able to connect with people. As an introvert myself, I can certainly relate to that feeling. But Matt, who also identifies as introverted, shared a powerful truth: “You don't have a choice, unless you wanna go spend three weeks never talking to somebody, you have to start striking up a conversation.”

This isn't about becoming an extrovert overnight. It's about building a new kind of social muscle. When you're traveling solo, the stakes are lower. You're likely never going to see these people again, so the pressure to impress or be perfect is gone. This freedom allows you to be more authentic and open. Matt suggests a few simple ways to meet people:

• Shared Group Activities: Think bar crawls, food tours, walking tours, or even finding local sports leagues (like pickleball or volleyball, as Matt's friends do). These activities provide a built-in common interest, making conversations flow naturally.
• Hostel Bars: Even if you prefer not to stay in a hostel, their bars are often vibrant social hubs where travelers gather. It's a low-pressure environment to strike up a chat.
• Meetup.com: If you have a specific interest, like jazz music, search for local meetups. You'll find like-minded individuals and an instant topic of conversation. As Matt says, “You at least have something you can talk about.” This initial common ground can quickly lead to deeper connections.

These aren't just tips for travelers; they're life lessons. Learning to initiate conversations, even when it feels uncomfortable, is a skill that serves you well in every aspect of life. It builds resilience and expands your social circle in unexpected ways.

The Art of “One Day, One Thing”

One of the most common pitfalls for remote workers or entrepreneurs who travel is trying to juggle work and sightseeing simultaneously. Matt has a strong, counterintuitive piece of advice: “Don't do it.” He explains that one will always suffer. If you try to work in the mornings and sightsee in the afternoons, you'll constantly feel stressed, either about work you're not doing or sights you're missing.

His solution is simple yet profound: dedicate entire days. “You have to be like, this is the workday and this is a travel day. You can only do one thing at one time.” This means if you're in Paris for a week, you might dedicate three days entirely to work and four days entirely to exploring. Or, if you have the luxury, extend your trip. As Matt points out, if you can work remotely, you can work anywhere. So why not go for a month? This allows you to truly immerse yourself in the rhythm of the city, find local coffee shops, and experience life like a local, without the constant pressure of splitting your attention.

This “one day, one thing” philosophy isn't just for travel; it's a powerful principle for productivity in general. Multitasking, as Matt reminds us, is a myth. Our brains aren't designed for it. By focusing on one task or one experience at a time, whether it's a work project or exploring a new neighborhood, you achieve deeper engagement and greater satisfaction.

Uncovering Local Gems: Beyond the Tourist Traps

When you're traveling solo, you have the freedom to deviate from the well-trodden path. You're not beholden to anyone else's itinerary, which opens up a world of authentic experiences. Matt shared a brilliant hack for finding these hidden gems: the local tourism board.

While most people skip it, he sees it as an “underrated resource.” The staff are locals whose job it is to help you have a good time. Instead of asking “What should I do?” (which will get you the usual tourist spots), ask “What do you do?” This simple shift in questioning can lead you to local markets, hidden restaurants, and unique experiences that most tourists never discover.

Another powerful strategy is to leverage local food or travel bloggers and Instagram accounts. As Matt explains, “Every city has their own version of that.” These feeds are curated for locals, offering insights into new restaurant openings, cool events, and offbeat attractions. It's the “who, not how” principle in action: instead of trying to figure it all out yourself, find the people who already have the answers.

The Unseen Benefits: Reading People and Building Resilience

Beyond the practical skills, solo travel cultivates a deeper understanding of human nature. Matt notes that you “get better at reading people” because you're exposed to more facial expressions and diverse situations. This heightened awareness helps you discern intentions and navigate social dynamics with greater ease.

It also builds incredible resilience. When things go wrong—and they will—you're the one who has to solve the problem. This might mean navigating a language barrier, finding an alternative route, or simply dealing with unexpected delays. Each challenge overcome strengthens your ability to adapt and problem-solve, making you more confident in your capacity to handle whatever life throws your way.

Your Next Adventure Starts Now

Solo travel isn't just for the adventurous few; it's for anyone ready to embrace personal growth and discover their own strength. It's about shedding the expectations of others and truly listening to your own inner compass. Whether it's a weekend trip to a nearby city or a month-long journey across a continent, the lessons you learn about yourself will be invaluable.

So, what's holding you back? Perhaps it's time to book that ticket, pack that bag, and embark on an adventure that will not only show you the world, but also show you the incredible person you are becoming.

This episode breaks down a simple, practical framework to help you identify what’s no longer serving you—and let it go with confidence. You’ll learn how to spot the hidden red flags, ask better questions about your commitments, and make small, low-risk changes that free up time, clarity, and momentum for what actually matters.

Ready to tackle bigger problems? Go to Claude.ai/tps.

Visit www.asianefficiency.com for more productivity tips and tactics.

Links:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Claude
• 25X Productivity Coaching

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

We've all been there. That feeling of being perpetually busy, constantly chasing the next deadline, the next achievement. You push yourself, you hustle, you optimize every minute, only to find yourself at the end of the day, week, or even year, feeling utterly drained, perhaps even a little lost. You might think you're being productive, but what if all that relentless striving is actually holding you back from true effectiveness and genuine well-being?

It's a common trap, one I've certainly fallen into. I remember a time, not so long ago, when I was on stage, ready to deliver a talk to a hundred people. My heart was pounding, my voice trembled, and a cold sweat broke out on my neck. I blamed it on the flu, but deep down, I knew it was anxiety. I pushed through, relying on autopilot, but the experience left me shaken. Lying in my hotel room afterward, I realized something profound: I wasn't in a good place. And that's when I knew I needed to find a way to bring more calm into my life.

This wasn't about abandoning productivity. Far from it. It was about understanding that true productivity isn't just about doing more, faster. It's about cultivating a state of mind that allows you to do your best work, to be present, and to find meaning in what you do. It's about recognizing that sometimes, the most productive thing you can do is to slow down, breathe, and find your calm.

The Hidden Cost of Constant Hustle: Why Anxiety and Burnout Cap Your Potential

We live in a world that often glorifies busyness. The more packed your calendar, the more emails in your inbox, the more projects on your plate, the more “successful” you're perceived to be. But this relentless pursuit of more comes at a cost. As Chris Bailey, a productivity expert and author, points out, anxiety and burnout are silent saboteurs that significantly compromise our cognitive performance without us even realizing it.

Think about it: if you had to solve a complex problem or give an important presentation while feeling overwhelmed and stressed, how well would you perform? Probably not your best. Your mind would be racing, your focus scattered. Now imagine tackling that same task with a calm, clear mind. The difference is palpable. Anxiety, like a thick fog, obscures your mental clarity, making even simple tasks feel monumental. It's like trying to multiply 12 by 24 in your head five minutes before a big talk versus doing it after a relaxing walk and a warm drink. The calm mind wins every time.

Burnout, often mistaken for mere exhaustion, is a deeper, more insidious phenomenon. It's characterized by three key components:

• Exhaustion: A state of total physical and mental depletion.
• Cynicism: A pervasive negativity towards your work and its purpose.
• Inefficacy: A profound feeling that your efforts make no difference, no matter how hard you try.

If you're experiencing all three, you're likely burnt out. But even if you're just feeling exhausted and cynical, you're well on your way. The opposite of burnout is engagement, and when you're burnt out, your capacity for productivity plummets. It's a spectrum, and the more chronic stress you face, especially from work, the further you slide towards burnout. This isn't just about feeling tired; it's about a fundamental erosion of your ability to perform and find satisfaction in your work.

The Six Dimensions of Work-Related Stress: A Blueprint for Engagement

Understanding the root causes of burnout is the first step towards cultivating calm and boosting your productivity. Chris highlights six key variables in our work that, depending on the level of chronic stress they induce, can either push us towards burnout or pull us towards engagement:

1. Workload: When your workload consistently exceeds your capacity, burnout is a real risk. It's not just about the quantity of tasks, but the relentless pressure to keep up.
2. Lack of Control: The less autonomy you have over what, when, and how you do your work, the more likely you are to feel disengaged and burnt out. While some roles offer more autonomy than others, there are always ways to find pockets of control.
3. Insufficient Reward: This isn't just about money. It's about feeling financially compensated, socially recognized, and valued for your contributions. When your efforts go unnoticed or unappreciated, it's a fast track to cynicism.
4. Community: Humans are social creatures. Feeling disconnected from your colleagues or lacking a sense of belonging at work significantly impacts your engagement. Strong social connections are a critical ingredient for a positive work environment.
5. Fairness: How work is assigned, how rewards are distributed, and how decisions are made all contribute to a sense of fairness. Perceived unfairness can breed resentment and disengagement.
6. Values: This is perhaps the most critical factor. When your work aligns with your deeper values, it becomes a source of meaning and motivation. When there's a disconnect, even if you're achieving a lot, it can feel hollow and unfulfilling.

By regularly assessing these six areas, you can gain valuable insights into your own well-being and identify areas where you can make adjustments. It's about being proactive, not reactive, in managing your stress and cultivating a more sustainable approach to work.

Reclaiming Control: Finding Autonomy in a Demanding World

Many of us feel like we have little control over our schedules, especially in corporate environments. I've heard it countless times: “My calendar is a wall of meetings!” And it's true, not everyone has the luxury of completely dictating their work hours. But even within demanding structures, there are ways to reclaim a sense of autonomy.

Chris emphasizes that we need to take the productivity advice that works for us and leave the rest. Not every tip will apply to every situation, and that's okay. The key is to understand your own autonomy spectrum. If you're in a role with limited control, focus on what you can control. This might involve:

• Defining “Productivity Hours”: Even if your day is packed, identify specific blocks of time where you can focus on deep work, free from distractions. Protect these blocks fiercely.
• Setting Boundaries: Learn to say no, or at least “not right now.” Disconnect after work hours. Resist the urge to constantly check emails or notifications. These boundaries, even small ones, create a sense of control and prevent constant overstimulation.
• Relating Differently to Your Work: Sometimes, it's not about changing your circumstances, but changing your perspective. How you relate to your work can be as important as the work itself. By consciously choosing to approach tasks with a calmer, more intentional mindset, you can reduce the perceived stress.

It's about working within the systems that exist, but also pushing and molding them for your collective benefit. Even small tweaks can lead to massive shifts in your sense of control and overall well-being.

The Power of the Pause: Embracing Stimulation Fasts and Analog Living

In our hyper-connected world, we're constantly bombarded with digital stimulation. Every notification, every new email, every social media scroll triggers a hit of dopamine, creating a craving for more. This constant seeking of novelty, while seemingly harmless, can actually diminish our capacity for focus and presence. It's like being in Times Square all the time; eventually, nothing feels truly novel.

Chris advocates for what he calls “stimulation fasts” – periods where you intentionally reduce your exposure to digital novelty. This isn't about completely cutting off from the world, but rather about creating boundaries around those “empty hits” of mental stimulation. It could mean:

• Weeding out digital distractions: Temporarily unsubscribe from news feeds, limit social media, or avoid binge-watching shows.
• Substituting with analog activities: Instead of scrolling, play an instrument, go for a walk in nature, spend quality time with loved ones, or engage in a hobby that brings you joy and calm.

The benefits are remarkable. When you lower your mental stimulation, focus becomes easier, almost effortless. You'll find fewer impulses to check your phone or refresh your email. This cultivated calm allows you to engage more deeply with meaningful tasks, leading to greater productivity and a stronger sense of purpose.

This concept extends to our choice of tools as well. While digital tools offer efficiency, analog tools often foster meaning. Think about writing a heartfelt letter by hand versus sending a quick email. The analog experience, though slower, often carries more weight and intention. It's about finding the right balance: use digital for efficiency, and analog for meaning. Deliberately choose when to embrace the slower, more present experience of analog living.

Your Path to Thoughtful Productivity

True productivity isn't about burning out in the pursuit of endless accomplishment. It's about cultivating a thoughtful approach that prioritizes your well-being, aligns with your values, and ultimately leads to more meaningful and sustainable results. It's about recognizing that calm isn't the absence of activity, but the presence of clarity and intention.

So, what's one small step you can take today to embrace thoughtful productivity? Perhaps it's identifying one digital distraction to reduce, or scheduling a short “stimulation fast” for yourself. Maybe it's reflecting on which of the six dimensions of work-related stress are most impacting you, and brainstorming one small way to address it. Or perhaps it's simply taking a few moments to breathe deeply and find a moment of calm amidst the busyness.

Upgrade your denim game with Rag & Bone!. Get 20% off sitewide with code TPS at www.rag-bone.com #ragandbonepod

Ready to tackle bigger problems? Go to Claude.ai/tps.

Make the switch! MINTMOBILE.com/PRODUCTIVITY.

Visit asianefficiency.com for more productivity tips and tactics.

Cheat Sheet:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Top 3 Productivity Resources [1:50]
• The Step Most People Skip Before Breaking Anything Down [5:16]
• Why Your 3 Available Hours Won't Fit a 20-Hour Plan [9:47]
• The Goal-Setting Framework Everyone Knows, but Nobody Uses [16:17]
• A Project Management Trick That Works for Personal Goals Too [17:43]
• He Asked AI to Break It Down — Then Said “Isn't That Obvious?” [21:41]
• The Daily Number That Predicts Whether You'll Hit Your Goal [22:51]
• Why Everything Feels Obvious After Someone Else Figures It Out [25:29]
• The One Daily Habit That Creates Unstoppable Momentum [28:25]
• What's Quietly Sabotaging Your Most Important Work [30:09]
• The 10-80-10 Rule That Redefines Your Role at Work [31:50]
• When You Realize You're the Weakest Link in Your Own System [33:50]
• The Paradox: Why Cheaper AI Means More Work, Not Less [36:32]
• The Hidden Tasks AI Just Made Worth Doing [40:24]

Links:

• Rag & Bone
• Claude
• Mint Mobile
• Sublime Text — Text editor with excellent color coding and support for markdown and configuration files
• Granola — AI meeting notes app that records and transcribes meetings without joining as a bot
• Jump Desktop — Remote desktop access tool for securely connecting to multiple computers
• The Goal: A Process of Ongoing Improvement — Eliyahu Goldratt's foundational book on Theory of Constraints
• 25X Productivity Coaching
• Laura Roeder Tweet
• Article about Garfield AI law firm
• Wispr Flow
• Productivity Academy
• The Productivity Show on Youtube
• The Productivity Show Youtube Clips
• Become a member of TPS+
• Take the Productivity Quiz

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

Ever felt like you're drowning in information, meticulously taking notes, highlighting, and organizing, only to realize you're not actually getting any smarter? You're not alone. It's a common trap in our information-rich world, where the sheer volume of data can lead to what I call “data paralysis.”

I recently came across an article by journalist Casey Newton titled “Why Note-Taking Apps Don't Make Us Smarter.” The headline immediately grabbed me, and as I read, I found myself nodding along. Casey described a journey many of us have been on: trying out countless note-taking apps, always searching for that one perfect tool that will unlock our intellectual potential. He tried everything from Rome to Obsidian, and even Mem, one of the first AI-powered note-taking apps. Yet, despite his diligent note-taking, he realized he wasn't becoming smarter; he was becoming more overwhelmed.

This isn't just about notes. Think about farmers today. They have access to an incredible amount of data about weather, precipitation, soil conditions, and more. You'd think this would make their lives easier, right? But often, they're overwhelmed by the sheer volume of information, making it difficult to actually use it effectively. It's the same with personal health trackers like the Apple Watch or Oura Ring. While I love my Oura Ring, and I'm obsessed with tracking my metrics, even I only focus on two or three key data points out of the dozens available. The rest just become noise.

So, if more notes and more data don't automatically lead to more intelligence, what's the missing piece? It's not the tools themselves, but what we do with the information they hold.

The AI Promise and Its Pitfalls

When AI started making waves, many of us, including Casey, thought, “Aha! This is it. AI will finally help us synthesize all this information and make us smarter.” Imagine chatting with your notes, asking an AI to summarize key points, or even generate action items from your meeting transcripts. Google's new NotebookLM, for example, aims to do just that, allowing you to interact with your own notes and learn from them. This is a direct response to tools like Microsoft Co-pilot, which integrates AI into the entire Microsoft 365 suite, enabling users to interact with emails, documents, and data through a chat interface.

These AI tools are incredibly powerful, and the potential is immense. You could ask Co-pilot to summarize a client email, or generate an email based on meeting notes. The promise is to make us more productive and efficient. But here's the catch: AI, as it stands today, isn't always accurate. If you've used ChatGPT, you know it can sometimes “hallucinate” or provide incorrect information. This means you still have to review and verify the AI's output, which, in a way, defeats the purpose of having it synthesize your notes in the first place.

So, while AI can be a fantastic assistant, it's not a replacement for critical thinking and synthesis. The real work of becoming smarter still falls on us.

The Unsung Hero: Your Weekly Review

Casey Newton came to a similar conclusion: no note-taking app or AI tool will make you smarter unless you actively engage with the information. You have to spend time thinking, synthesizing, and connecting the dots yourself. That's where true insights emerge. And this is where a consistent routine, like a weekly review, becomes your secret weapon.

I always tell my coaching clients that a weekly review is non-negotiable. Every Sunday, I dedicate time to go through all my notes, everything I've captured throughout the week. This isn't just about organizing; it's about processing, synthesizing, and turning information into actionable steps. It's where I connect ideas, identify patterns, and transform raw data into valuable insights. Without this dedicated time for reflection and processing, all the notes you take, no matter how meticulously organized or AI-summarized, will have little value.

Think about it: if you're just capturing information without ever reviewing it, it's like collecting ingredients for a meal but never actually cooking. The potential is there, but it remains unrealized. The power of your notes comes not from their existence, but from your interaction with them.

Beyond Notes: The Power of Routine and Purpose

This concept extends beyond note-taking. I recently encountered a Reddit post from a 26-year-old woman who was struggling to get out of bed, often waking up at 2 PM. She was unemployed and felt lost, even resorting to an alarm app that required scanning a QR code to turn off. While the alarm app is a creative solution, her core problem wasn't a lack of alarms; it was a lack of routine and purpose.

Many people face this challenge. When there's no structure, no clear routine, it's easy to become reactive, letting the day dictate your actions rather than proactively shaping it. This is why a “rise ritual” or morning routine is so crucial. It provides a sense of structure and accountability. Professional athletes, for example, have highly structured lives with clear schedules for training, practice, media, and games. But when they retire, many struggle because that external structure is gone. They suddenly have to create their own routine, which can be incredibly challenging after years of being told what to do.

For this woman, and for anyone feeling adrift, the first step is to implement a routine. Start with a rise ritual. It doesn't matter if you wake up at 6 AM or 2 PM; the key is to have a consistent set of actions you take as soon as you wake up. Open the curtains, get some sunlight, and then engage in a planned activity. This creates a sense of momentum and purpose.

Another powerful strategy is to define a “daily target.” Before you start your day, know what a productive day looks like for you. For the unemployed woman, it could be submitting resumes to 15 companies, or focusing on finding a job within the next 30 days. When you have a clear goal, every day becomes a focal point, guiding your actions and giving you something to work towards. Without a target, you're just drifting.

The “Busy Person” Paradox

It might sound counterintuitive, but often, the busiest people are the most productive. I've noticed this in my own life: when I have a packed schedule, I become incredibly efficient at squeezing in tasks and focusing on what truly matters. But during periods with less structure, even if I have more free time, my productivity can actually dip. The external pressure and clear deadlines of a busy schedule often force us to be more intentional with our time and energy.

This isn't to say you need to be constantly overwhelmed to be productive. Rather, it highlights the importance of self-imposed structure and clear objectives. If you don't have external deadlines, create your own. If you don't have a boss telling you what to do, be your own boss.

The Takeaway: Engage, Synthesize, Act

In a world overflowing with information and increasingly sophisticated tools, the real differentiator isn't how much you consume or how many apps you use. It's how actively you engage with that information, how consistently you synthesize it, and how diligently you translate it into action. Your notes, your data, and even the most advanced AI are merely raw materials. You are the architect of your own intelligence and productivity. So, stop just collecting, and start creating.

Try Notion Custom Agents at notion.com/tps.

Links:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Notion
• 25X Productivity Coaching
• The Most Underrated Productivity Technique
• First Things First by Steven Covey

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

The digital world is overflowing with tools promising to boost your productivity. From project management software to fancy to-do list apps, it feels like there's a new must-have solution every other day. But what about something as fundamental as note-taking? We all do it, whether it's jotting down a quick thought, capturing meeting minutes, or brainstorming a new idea. Yet, many of us still struggle to find a note-taking system that truly works for us. It's like trying to find the perfect pair of shoes… you know it's out there, but the sheer number of options can be overwhelming. You end up with a closet full of shoes you barely wear, and your feet still hurt. Sound familiar?

Today, we're diving deep into the world of note-taking apps. We're not just talking about basic digital scratchpads; we're exploring tools that can transform how you capture, organize, and leverage your ideas. We'll cut through the noise and highlight three distinct approaches to note-taking, helping you find the “perfect fit” for your unique needs. Because when your note-taking system clicks, it's not just about being organized… it's about unlocking new levels of clarity, creativity, and ultimately, productivity.

The Need for Speed: Why Quick Capture is King

Think about those moments when an idea strikes, or you hear a crucial piece of information. It could be in a meeting, during a casual conversation, or even while you're out for a walk. If your note-taking app isn't fast and easy to use, that valuable insight might vanish before you can even open the app. It's like trying to catch a fleeting thought with a slow net… by the time you're ready, it's gone. This is why the speed and ease of use are paramount when choosing a note-taking tool. If there's too much friction, you'll simply stop using it.

Beyond speed, seamless synchronization across all your devices is a non-negotiable. Imagine taking notes on your phone during a commute, then wanting to pick up exactly where you left off on your laptop at the office. Without reliable sync, you're stuck emailing notes to yourself or, worse, retyping them. And finally, a powerful search function is your best friend. Even if you're not a meticulous note-namer, you should be able to find what you need with just a few keywords. It's like having a personal librarian who knows exactly where everything is, even if you just tossed it onto a pile.

The Familiar Comfort of Classic Evernote: Nimbus Note

For those who fondly remember the golden age of Evernote, or for anyone seeking a robust yet straightforward note-taking experience, Nimbus Note is a strong contender. It offers a familiar interface and feature set that mirrors the classic Evernote experience. Think of it as a comfortable, well-worn jacket that still looks great and keeps you warm. It's cross-platform, available on web, Mac, Windows, iOS, and Android, ensuring your notes are accessible wherever you are.

Nimbus Note allows you to create well-structured notes with tables, tasks, and lists. But it goes a step further than old Evernote by offering a true folder structure, giving you more control over your organization. And just like Evernote, it boasts an excellent web clipper, a feature I personally rely on heavily. This means you can easily save articles, recipes, or any web content directly into your notes, keeping all your research and inspiration in one place. While I haven't personally used Nimbus Note extensively, its resemblance to the beloved classic Evernote makes it a compelling option for anyone looking for a reliable and feature-rich note-taking solution without the complexity of some newer tools.

The All-Encompassing Powerhouse: Notion

If you're looking for a note-taking app that can do it all, and then some, Notion is likely already on your radar. It's not just a note-taking app; it's a versatile workspace that can manage your entire life and business. Imagine a Swiss Army knife for your digital world… it has a tool for every possible need. Notion has exploded in popularity, and for good reason. It combines note-taking with database features, project management, and even website creation, all without requiring any coding skills.

One of Notion's biggest strengths is its flexibility. You can use it for personal journaling, managing complex team projects, or even building a personal knowledge base. The possibilities are truly endless. While it's primarily web-based, requiring an internet connection for full functionality, its robust feature set and active community make it a go-to for many. The template feature, in particular, has been a game-changer. Need a habit tracker? A content calendar? A CRM? Chances are, someone has already created a Notion template for it, allowing you to get started quickly and customize it to your heart's content. This vibrant ecosystem of shared templates and workflows has created a powerful network effect, making Notion incredibly accessible and adaptable.

The Knowledge Ecosystem Builder: Obsidian

For those who crave a deeper, more interconnected approach to their notes, Obsidian offers a unique and powerful solution. It's not just about taking notes; it's about building your own personal knowledge management (PKM) system. Think of it as constructing a vast, interconnected library of your thoughts and ideas, where every piece of information can be linked to another. This allows for a more organic and insightful way of exploring your knowledge.

Obsidian's true power lies in its ability to link notes together, creating a web of interconnected ideas. This can lead to unexpected insights and a deeper understanding of your own thoughts. It's available across multiple platforms, including Mac, Windows, Linux, iOS, and Android. A key differentiator for Obsidian is that it stores your notes in plain markdown text. This means your notes are not locked into a proprietary format, giving you complete control and ensuring future accessibility. If Obsidian were to ever cease to exist, your notes would still be easily readable and transferable. While the learning curve can be a bit steeper than other apps, the potential for creating a truly personalized and powerful knowledge ecosystem is immense. If the idea of a connected web of your own thoughts excites you, Obsidian is definitely worth exploring.

Finding Your Perfect Fit

So, how do you choose the right note-taking app for you? It's like trying on different pairs of shoes. The best way to know if it fits is to try it on. Pick the app that resonates most with you from these three options, or even one of the built-in options like Apple Notes or Google Keep. Install it, download it, and give it a solid two-week test drive. Don't feel pressured to commit for life. If it doesn't feel right, you can always switch back or try another. The goal is to find a system that reduces friction, enhances your ability to capture and recall information, and ultimately, helps you be more productive.

Sign up for a $1/month trial period at shopify.com/tps.

Try Notion Custom Agents at notion.com/tps.

Masterclass.com/TPS for an additional 15% off any annual membership.

Visit asianefficiency.com for more productivity tips and tactics.

Cheat Sheet:

Become a member of TPS+ and get ad-free episodes a week before anyone else with other great bonuses like the famous “One Tweak A Week” shirt.

• Top Three Productivity Resources [1:56]
• Running 300 Lead Enrichments Autonomously [5:17]
• Why Real-Time Search Still Matters [9:43]
• Finding the Right AI Tool for the Job [11:23]
• Securing Your AI Browser Sessions [17:42]
• Chrome Plugin Workflows That Run Overnight [19:43]
• Voice Mode Goes Everywhere [22:48]
• The Dictation Upgrade Changing Everything [26:20]
• Sounding Human in AI-Generated Writing [29:19]
• Context Files Transform Voice Inputs [30:31]
• Closing the AI Feedback Loop [32:11]
• Skills Beat Generic AI Prompts [34:16]
• Learning AI Through Structured Frameworks [38:07]
• From Beginner to Advanced in Three Stages [41:05]

Links:

• Shopify
• Notion
• Masterclass
• 25X Productivity Coaching
• 4-Day AI Sprint
• AI Workshop in Austin TX
• Perplexity Computer — Autonomous AI agent system for complex task execution
• Airtable — AI-native database platform with native integrations for AI agents
• Gamma — AI presentation and proposal builder with Claude integration
• Google NotebookLM — AI research tool for analyzing documents and generating audio overviews
• ChatGPT — OpenAI's conversational AI
• Claude — Anthropic's AI assistant
• Google Gemini — Google's multimodal AI assistant for images and video
• Grok — xAI's real-time search and reasoning AI
• Claude for Chrome — Browser extension enabling Claude to read, click, and automate web tasks
• Exa AI — AI-powered search and scraper tool for finding people and data; used in recruiting workflows
• Manus AI — Autonomous AI agent that executes complex tasks end-to-end; recently acquired by Meta
• Wispr Flow — AI-powered voice dictation app removing filler words and auto-editing
• OpenClaw — Automation and AI agent platform
• Rick Mulready's YouTube Channel — Recommended YouTuber for accessible AI education
• Productivity Academy
• The Productivity Show on Youtube
• The Productivity Show Youtube Clips
• Become a member of TPS+
• Take the Productivity Quiz

If you enjoyed this episode, follow the podcast on Apple Podcasts, Spotify, Stitcher, Overcast, Pocket Casts or your favorite podcast player. It’s easy, you’ll get new episodes automatically, and it also helps the show. You can also leave a review!

I remember a time when my calendar felt like a relentless taskmaster, a never-ending list of obligations that dictated my every move. Each entry was a reminder of something I had to do, a commitment I couldn't escape. It was a constant battle against the clock, a race to tick off boxes, and honestly, it left me feeling drained and perpetually behind. Sound familiar? Many of us, especially those driven to achieve, fall into this trap. We see productivity as a game of efficiency, cramming more into less time, and our calendars become a testament to our busyness. But what if I told you that this very tool, often seen as a symbol of our overwhelm, holds the key to unlocking a life of true creation and purpose?

This isn't about adding more to your plate or finding new hacks to squeeze in extra hours. It's about a fundamental shift in how you view your time and, more importantly, how you interact with your life. My friend and master coach, Josselyne Herman Saccio, has a profound way of looking at this. She challenges the very notion of “being busy” and introduces a powerful concept: being unmessable with.

Unmessable With: Beyond Toughness, Towards Creation

Josselyne defines “unmessable with” not as being tough or impervious to feelings, but as the ability to stay in a mode of creation, no matter what life throws your way. It's about acting from your vision and purpose, rather than reacting to circumstances. Think about it: when you're constantly reacting, you're letting external forces dictate your actions. But when you're creating, you're in control, shaping your reality from the inside out.

One of the biggest roadblocks to this creative state, Josselyne points out, is the disempowering conversations we have with ourselves. Those limiting beliefs we picked up as kids, like needing to be the “best at everything” to be accepted, can hold us back. Josselyne's own story is a perfect example. She gave up on her dream of being a singer because of a childhood belief. But once she dismantled that belief, she landed a record deal within three months. That's the power of becoming unmessable with.

Your Calendar: More Than Just a To-Do List

So, how do you start cultivating this “unmessable with” mindset, especially when your to-do list feels endless? It begins with understanding the context of your actions. Josselyne illustrates this beautifully: a finger can be a finger, a number one, or an upward direction, depending on the context. Similarly, your tasks take on different meanings based on the context you create for them.

Most of us react to our to-do lists. We just do things because they're there. But Josselyne suggests a radical shift: create a context for every item in your calendar. Instead of just “go to the gym,” frame it as “creating vitality by going to the gym.” This simple change transforms a chore into an act of creation. It's about understanding why you're doing something, not just what you're doing.

This isn't a magic pill. It's a muscle you need to build. Start by noticing when you're being reactive. Are you overthinking? Do you feel a tightness in your chest? These are your red flags. When you catch them, take a breath, and consciously shift into creation mode. Pre-create your context. Write it down. Put it on a sticky note. Tattoo it on your arm if you have to! The goal is to have a clear purpose for your actions, so you're not relying on your brain to figure it out in the heat of the moment.

Overcoming Creative Blocks and Imposter Syndrome

What about creative blocks? Josselyne, a multi-talented artist and author, schedules her creative time. If it's not in the calendar, it's less likely to happen. And for those moments when the well feels dry, she suggests using AI tools to spark ideas. It's not about letting AI do the work, but using it as a springboard to get your own creative juices flowing.

And imposter syndrome? Josselyne has a powerful antidote: remind yourself of your results. When she left Landmark after 30 years, she had to build a new business and social media presence from scratch. She felt that familiar “who am I to do this?” But by writing down every result she had ever produced, she realized the depth of her accomplishments. It's hard to argue with facts. So, if you're feeling like an imposter, list out your wins. It's a powerful way to reconnect with your capabilities.

Living an Unmessable Life

Josselyne's daily life is a testament to this philosophy. She doesn't need motivation because she has the power of her word. Her word is her wand, creating her reality. She schedules everything, from hot yoga to intimate time with her husband, ensuring that what matters gets done. She even works with her assistants using tools like Asana, but she's not a slave to the tools. She adapts what works for her, like the time-energy audit, and discards what doesn't, like never touching her calendar.

So, if there's one key takeaway, it's this: start with your calendar. Use it as a canvas to not only create what you're doing, but why you're doing it. Infuse every entry with context and purpose. It's not just a to-do list; it's the creation of your life. And when you do that, you become truly unmessable with.