Something pretty major just happened in the internet world, and I’ve got thoughts.

On July 1st, Cloudflare announced they’re now blocking all AI crawlers by default—unless a site explicitly says, “Come on in.” As someone who uses Cloudflare across both my personal projects and our work at Augusto, I’m giving this move a solid, two-thumbs-up, slow clap.

What’s the Big Deal?

Let’s break it down: right now, AI bots roam the internet like overly curious raccoons—scraping up content, training models, and sometimes spitting back that same info without ever pointing back to the source.

• No credit.
• No link.
• No traffic
• No compensation.

Meanwhile, the people and teams who actually wrote the blog posts, created the tools, or built the guides? Left wondering why their traffic dropped while ChatGPT suddenly sounds suspiciously like their site.

That’s the problem Cloudflare just decided to confront head-on.

What Did Cloudflare Do?

Cloudflare, which powers roughly 1 in 5 websites on the internet (yes, really), is now blocking AI bots like GPTBot, ClaudeBot, and others by default on their Free, Pro, and Business plans.

Want to train your next LLM? Great—go ask for permission first. Or maybe even talk about compensation. (This will be interesting)

This flips the script. Instead of publishers needing to jump through hoops to opt out, AI companies now have to work to opt in.

Why I’m Into This

I’ve been a Cloudflare fan for years. I use it for everything from DNS to WAFs to performance tuning. But this move feels bigger than a feature toggle—it’s a stand.

As someone who builds digital products and works closely with creators, devs, and businesses trying to grow, this is huge. It sends a clear message: your content has value.

I’m all for AI. (You’ve probably heard me talk way too much about automation, agents, and LLM workflows.) But I’m also for fairness. If AI is built on the backs of real people’s writing, research, or creativity—let’s at least give those people a seat at the table.

Or hey, maybe even a royalty check.

Here is an image of Cloudflare, an AI crawler on this site

What This Means for All of Us

This changes how we think about:

• Content ownership – Who gets to decide how your work is used?
• Attribution and traffic – Should AI models send users back to the source?
• Value exchange – If AI tools profit off your work, should you be cut in?

We’re entering a phase where content licensing for AI might be a real, scalable business model. And that’s exciting—especially for creators, educators, and product companies who’ve been burned by the “take now, ask never” approach.

Part of me realizes that AI is now probably now the new search engine and SEO is changing. The question is where do you want your content or your information.

As a small business, I would like AI to analyze my content. If I were a larger syndicate for news or content, I would want to discuss its value.

Final Thought

I think Cloudflare just lit a match. This is a pivotal moment in how we balance innovation and respect on the internet.

The AI tools we’re building need great content to learn from. But we also need to protect the people who make that content.

So yeah—I am interested to see how this progresses.

AI is everywhere these days, but how much does it actually cost to use? If you’ve played with ChatGPT, Claude, or similar tools, you’ve probably seen terms like tokens and credits. Here’s a quick guide to what they mean—and how to avoid surprises on your bill.

What Are Tokens?

AI models don’t count words—they count tokens. A token might be a word, part of a word, or even just a character. For example, “Hello world!” is three tokens: “Hello”, “world”, and “!”. Both your input and the AI’s output use tokens, so longer conversations or big documents use more.

What Are Credits?

Credits are how you pay for AI use. Some platforms give you monthly credits, others charge by the number of tokens you use (like $0.002 per 1,000 tokens for GPT-4.1 input). More complex tasks—like generating images or analyzing audio—cost more credits because they use more computing power.

Model Costs: Not All AI Is Priced the Same

Different models have different strengths—and prices. Here’s a quick snapshot:

Pro tip: Use lighter models for routine jobs (like parsing invoices), and save the heavyweights for complex or creative work.

Watch Out for Hidden Costs

It’s easy to burn through credits without realizing it—especially if you’re debugging code, running automations, or just chatting a lot. Always keep an eye on your usage dashboard, set limits, and look for ways to optimize.

You are out of Credits

I’ve often found myself running out of tokens on various tools. It all began with ChatGPT when I was using the free version—I was conducting research and quickly used up my tokens. Now, I am a subscriber. This trend continued when I started using CoPilot in my development work and later explored tools like Loveable.dev and Bold.new.

I now have multiple subscriptions, but I have trained myself to effectively utilize the credits and tokens in my development and workflow processes. How many of you have run out of tokens while in the middle of something you needed to complete?

This is may be a message you have seen :

Quick Tips to Save on AI

• Keep prompts short and clear.
• Pick the right model for your task.
• Cache results when possible.
• Audit your automations for unnecessary calls.
• Test with limits to avoid runaway costs.

The Bottom Line

Tokens are what you use; credits are how you pay. The model and task you choose are your cost multipliers. Understanding these basics will help you get the most.

What is your experience with Tokens and Credits?

It turns out, I’ve been self-hosting software at home for over 20 years. What started with a bulky server humming in the corner has evolved into something far more compact and elegant. Today, my entire setup runs on a small Intel NUC and a NAS, both quietly sitting on a shelf. Despite the size change, the number of services I run has only grown.

🛠️ From Racks to Raspberry Pis

In the early 2000s, self-hosting meant hours of tinkering, huge hardware, and living with the constant hum of fans and spinning disks. But over time, cloud-native apps and tools like Docker changed everything. Now I can spin up a new app in seconds, isolate it from the rest of my environment, and tear it down just as easily.

Recently,  I ran a quick script to export a list of all the Docker containers I’m currently running—and wow, it’s grown! I realized just how central self-hosting has become in managing both my personal productivity and home automation.

LIST OF ALL OF MY DOCKER CONTAINERS: https://a.ugus.to/selfhostjim

🔧 Tools I Can’t Live Without

Here are a few of the services I self-host today that I consider absolutely essential:

1. Habit Tracker

Staying consistent is everything. I self-host a lightweight, no-frills habit tracker that helps me reinforce the routines I care about most—morning workouts, reading, and even making time to explore new tech.

2. n8n – Automation and AI

This is the beating heart of my daily workflow. I use n8n not just for basic task automation, but also to pull in AI-generated summaries, daily briefings, and reminders. It’s my own personal assistant, running 24/7, and I couldn’t imagine my day without it.

3. Plex

For media, nothing beats Plex. It gives me a unified, beautiful interface to stream all of my personal content—movies, music, and old recordings—with zero monthly fee and total control over how it’s organized.  We use this at our home for Live TV , TV Shows, and Movies.

4. Home Assistant

This is where the magic happens. Home Assistant helps me automate everything from lights and climate to alerts and energy usage. It’s integrated with everything—from Zigbee sensors to my smart thermostat—and lets me run my house exactly the way I want.

5.  OpenWebUI an Ollama

I run my own type of ChatGPT using Ollama (local LLM on my Mac M1) and also allows me to use AI offline.

🧪 Why I Still Self-Host

So why do I still self-host when so many cloud services are available?

• Ownership: I want full control over my data and how it’s used.

• Customization: I can tweak things in ways most SaaS tools don’t allow.

• Privacy: I know exactly where my data is, who can access it, and what it’s doing.

• Learning: Every self-hosted app is an excuse to learn something new.  Especially AI.

💬 What About You?

I shared a list of my running containers with the team at Augusto for fun—and now I’m curious:

What are you self-hosting? What tools do you rely on every day?
Whether it’s automation, media, personal productivity, or just cool experiments, I’d love to hear about it.

Let’s swap notes. 🚀

• Replit AI 
• Bold.new
• Loveable.dev
• V0.dev

Each tool takes a slightly different approach to AI-assisted development, but by using the same use case across all of them, I got a clearer sense of what each one offers, where it shines, and where it still needs work.

Why AI Dev Tools Matter

Before diving into the comparisons, let’s talk about the “why.” Traditional software development requires deep technical knowledge and time-consuming tasks like boilerplate coding, debugging, and documentation. AI development tools promise to:

• Accelerate development by generating code from simple prompts or UI sketches.
• Lower the barrier to entry for non-engineers or less technical team members.
• Free up developers to focus on problem-solving and creativity.
• Support rapid prototyping for startups and internal tools alike.

This evolution doesn’t replace developers—it enhances them.

Example App: A weather APP

To help evaluate these tools, I decided to use a simple weather app that takes an input of Name and Zip code.
The App utilized a third-party API (OpenWeather API) to help showcase how these tools can interface with other systems. I also asked the system to try to capture the user’s location from the browser or their mobile device.

Outcome and results

Look at the videos and let me know if you have feedback.

Overall, each tool is very powerful. I have a longer review where I will compare each and rate them, which will be posted on our Augustodigital.com blog.

Here is a link to videos for each AI dev tool as I create a weather application.

• Part 1- Bolt.new Video
• Part 2 – Loveable.dev Video
• Part 3 – V0.dev  Video
• Part 4 – Replit AI – Video

YouTube Playlist for the videos: https://cutt.ly/Jra08i5W

Automating Business Workflows

At work, n8n plays a critical role in handling various automation needs, including:

• Finance & Budgeting – Automating time tracking, budget updates, and financial reporting.
• Operations & Reporting – Streamlining operational processes and generating automated reports.
• API Integrations – Connecting different services and ensuring smooth data flow across platforms.
• ETL & Data Processing – Extracting, transforming, and loading (ETL) data for analytics and decision-making.

These automations save time, reduce errors, and improve efficiency, making n8n an invaluable asset to our business operations.

Personal Use: Organizing My Day & AI Workflows

Beyond work, I use n8n to automate various personal tasks:

• Daily Summaries & Task Management – Automatically compile emails with important updates and reminders.
• Weather Forecasts – I get daily weather reports sent directly to me. I also have weather and water temperature apps that I call from telegram or text to give me the water temperature in Holland and send me an image. (see image)
• Local AI & LLM Processes – Running automation for my self-hosted AI models and local LLMs.
• Alexa app: I have an Alexa app that uses n8n to run an API that provides weather and other information about security cameras
• Home Alarm and Security – N8N has automation and workflows that I use to track my security campers and home alarm notifications.

Lately, I’ve been diving deep into local LLMs, leveraging n8n as the engine to process and orchestrate AI workflows. My self-hosted setup includes Open WebUI and models like DeepSeek and Qwen2.5, allowing me to run my own AI chat engine.

My Self-Hosted n8n Setup

I run my n8n instance on Docker, making it easy to manage and scale. To ensure secure and remote access, I use Cloudflare Tunnels, which allow me to connect n8n with other applications safely.

Why You Should Try n8n

If you haven’t used n8n yet, I highly recommend giving it a try. It’s an open-source, low-code/no-code workflow automation tool that allows you to build powerful automations with minimal effort. Whether for business or personal projects, n8n helps you streamline tasks and focus on what really matters.

Check it out at n8n.io and start automating today!  If you want to chat on this topic or learn more, please contact me.

What are you building with N8N?

There’s so much to share about Cloudflare that it’s impossible to cover it all in just one blog post.

Why Cloudflare?  Cloudflare isn’t just a CDN (Content Delivery Network) —it’s an all-in-one powerhouse for internet security, performance, and reliability. Here’s why I rely on it every day:

Services I Use:

• Domain Registry
  • Cloudflare’s domain registration offers an excellent registration tool and great pricing—no hidden fees, just straightforward value.
• DNS Services
  • Lightning-fast DNS resolution keeps my domains always reachable.
  • DNSSEC protection shields against spoofing and hijacking attempts.
    
  • DNS proxy and dynamic DNS are great features.
• SSL
  • Automatic SSL/TLS encryption ensures secure connections across my websites.
  • One-click certificate management makes securing domains effortless.
• Analytics
  • Real-time insights give me a clear picture of traffic and performance.
• Security (Proxy)
  • Cloudflare’s Web Application Firewall (WAF) stops threats before they can cause harm.
  • DDoS protection at all layers ensures my sites stay online, even during attacks.
  • The Zero Trust security model strengthens access control and user authentication.
• Tunnels
  • Cloudflare Tunnels provide secure, private access to my services—no public exposure needed.   This is such a powerful tool.  I’ll be sharing a few posts on how I use this for home and work. 
• Bot Blocking
  • Advanced bot management stops bad actors in their tracks.
  • Prevents spam, scraping, and automated attacks without affecting real users.
• Cloudflare Pages – JAMstack
  • Pages is a fantastic service, I use this to deliver client demos and share AI products (more to come)
  • Serverless hosting with seamless JAMstack deployment.
  • Edge functions optimize performance and scale effortlessly.

Final Thoughts As someone who values security, speed, and efficiency, Cloudflare is an absolute game-changer. I find my self using Cloudflare for both work and home use.  Whether it’s domain management, security, or performance tuning, Cloudflare delivers. If you’re not using it yet, you’re missing out on the best internet services available.

Want to chat about how Cloudflare can enhance your setup? Let’s connect!

Many people I have talked with lately are asking questions about deepseek.  They have heard about it on th enews and ask what is it.

When talking to people I also mention that I have been running my AI models locally.  This has really allowed me to fine tune, train, and also experiment with AI.

So, I did a quick video to show how someone could use DeepSeek locally.  In this video I am running ollama and DeepSeek on a macbook and it allows me to keep all of my prompts and data local.
I have been working with AI in various aspects of my life, experimenting with different applications. I’ve changed the way I approach AI in strategy and business. One of the advantages of being in the technology field is not only using the technology but also teaching others about it.

Recently, many people I’ve spoken with have been asking about DeepSeek. They’ve heard about it in the news and want to know more about what it is.

When discussing this, I also mention that I’ve been running my AI models locally. This approach has allowed me to fine-tune, train, and experiment with AI more effectively.

To share my knowledge, I created a quick video demonstrating how someone can use DeepSeek locally. In this video, I run Ollama and DeepSeek on a MacBook, which enables me to keep all my prompts and data on my local machine.

I recently updated my main computer, a MacBook Pro, to a more current version. I appreciate how easy Apple makes the migration process. You reset the Mac, and then you migrate from one Mac to another.

Family & Focus

First and foremost, my family is doing well, and they remain my number one priority. Everything else comes second to them.

On the career side, my focus is on Augusto Digital, the company I co-founded with two close friends. It’s been an incredible journey building something meaningful with people I trust and respect.

Passion for Soccer & Technology

Some things never change—my love for soccer and technology is as strong as ever.

Soccer is more than just a game for me. It’s my stress relief, my mental health reset, and a way to keep both my mind and body in sync. No matter how busy life gets, I make time for it.

Then there’s technology, which never fails to excite me. It’s constantly evolving, and I love staying ahead of the curve.

What I’ve Been Working On

I’ve been keeping busy both professionally and personally, diving into:

Technology I’m Working On

1. Self-Hosting & Open Source – I’ve been a fan of self-hosting for a long time, running multiple systems and technologies locally. This passion for open-source tools has not only benefited my personal projects but has also shaped how we work at Augusto Digital. Tools like Docker, n8n, Appsmith, and others have become key solutions for both home and business.

2. Automation – I’m always trying to automate more aspects of my home. From Alexa to Home Assistant, I’ve set up scripts, AI routines, and IoT devices that make our home safer and more enjoyable. Nothing beats hearing my kids casually ask Alexa to turn on the lights like it’s second nature.

3. AI & Machine Learning – I’ve been experimenting with AI for both work and personal use. ChatGPT, Grammarly, and other tools help me communicate better and adjust to different contexts. At work, I’ve been diving into RAG (Retrieval-Augmented Generation) and fine-tuning models to enhance how we use AI in our apps. Lately, I’ve been experimenting with local AI agents like Ollama, which allow you to run private AI models and keep your data secure. Running Deepseek locally has also been a fascinating experience. If you’re interested in this, reach out—I’d love to chat.

4. n8n & Workflow Automation – I don’t think enough people realize how powerful n8n is. It’s an incredible tool for automation, and I use it extensively both at work and in my personal life. At Augusto Digital, n8n helps with time management, reporting, integrations, and even our Data Warehouse. Personally, I’ve set up automation for daily habits, tasks, weather updates, and more, making sure key information reaches me when I need it.

5. AI & Development – At Augusto Digital, I manage our software engineering team, and we have an incredible group of people. AI is becoming a key part of how we enhance our development process. GitHub Copilot has been a game-changer in reviewing, improving, and accelerating development. Beyond that, I’ve been experimenting with tools like Bolt.new, Loveable.dev, and Copilot Agents, which are pushing the boundaries of AI-assisted software creation. Over the past few months, I’ve been testing these tools to build prototypes and evaluate how far along these systems are in generating production-ready software. It’s an exciting time to be in development—seeing what’s possible with AI is truly incredible.

Lessons from the Last Few Years

One thing I’ve learned over the years is the importance of connecting with people who are smarter than me, experimenting with new technology, and sharing as much as possible. The more I learn, the more I realize how much there still is to explore.

I’ll try to be better about updating this space more often. But no promises—just a commitment to keep learning, building, and improving.

Until next time!

I recently had the need to create a small python script to run a process.  I quickly looked up a process I could utilize to read a Rest API then do some work based on that content.

I fired up Visual Studio code and wrote out the framework of the program I needed in about 10 min.  (NOTE: I have not used Python in a long time and it was fun to quickly put something together).

Once I was done writing the code I jumped into a term terminal and ran the program “python GetAPIData.py”.  It worked well – opened a file of records added to an array and displayed the line items and did a count.

I then added some code that referenced how to call an API and it used the “requests”, knowing I needed to run “pip install requests” before I ran the program, I was rewarded with an error.

 zsh: command not found: pip

After this error, I checked what python version was on my Mac (python –version) I then learned I was running python 2.7.

This was confusing as I had just run “brew install python” before starting the project.  Apparently, Mac comes with python 2.7 and when you install brew it does not update the path. I was able to run the app via command python3.  And install “requests” via pip3 requests.  But this was not the best solution for me.    I really just want to run python 3 all the time.

The solution I found was to alias Python to python3 in my profile.

I updated my .zshrc file with the following nano ~/.zshrc and added this to the last line “alias python=python”

After this, I had one more error.  I was using the print command incorrectly, python3 has the print function using () vs “”.

Here is what has changed:

print 'Hello, World!'
print('Hello, World!')

Lesson learned – Take a look at what version of the app you are using.  I spent more time figuring out how to update python on my Mac than I did writing my code.