• Digital sovereignty: Federal government withdraws from AI partnership with Mistral and SAP – “The [German] federal government is withdrawing from the partnership with Mistral and SAP. The decision fuels doubts about Europe’s path to digital sovereignty.”
• Home Depot and Lowe’s share a common goal for their AI assistants, but their strategies diverge
• Cabinet recognizes: Cloud Act can collide with AVG and give access to Dutch healthcare data – [Translated:] Furthermore, there is no insight into American access to Dutch patient data. The US government can secretly request our healthcare data. The government does not know if Dutch care data has ever been requested. The lack of reports is certainly not proof that such requests do not take place.”
• Not enough nerds left Twitter, it’s still relevant – “Since that blog post, Federico Viticci and John Voorhees returned to X. Other people started posting to X who had never even posted to Twitter before, like Jensen Huang and Jeffrey Katzenberg. While I still believe X is doomed in the long run — closed, rudderless, and toxic to many — the AI community has held on to a remnant of the old Twitter, and in fact that community is growing, apparently immune to the decay of a platform that has no soul.”
• Tiny5 – “Tiny5 is a family of compact 5-pixel variable fonts that capture the essence of 1980s – 90s digital minimalism. Inspired by the graphing calculators and pocket gadgets of the era, they distill letterforms to their absolute essentials – proving that even at just five pixels tall, clarity, charm, and personality can thrive together.”
• PC Market Woes Continue: Shipments Fall 20.1% in Q3 2026 as Pull-In Hangover and Supply Constraints Bite – IDC – ”Worldwide PC shipments fell 20.1% year over year in Q3 2026 to 62.7 million units, according to preliminary results from IDC Research (IDC), the second consecutive decline and a deeper one than the 3.8% drop in Q2. ”
• Anime, etc., now completely normal, even for the olds – “A new NRG survey commissioned by anime streamer Crunchyroll found that in the United States, the percentage of consumers aged 13 to 54 who watch Japanese animation at least weekly increased to 41 percent this year, up from 39 percent. The percentage of daily watchers also increased to 21 percent from 13 percent in 2025. The idea that people age out of anime was rattled by the survey, with 58 percent of viewers aged 45 to 54 indicating that their fandom had intensified.”
• AI coding has made CI a bottleneck, so we reworked ours to keep up – “Some of our earliest gains required almost no optimization of CI itself. Moving our workloads off GitHub Actions to third-party runners with faster CPUs, higher-performance storage, and better cache infrastructure gave us faster machines to run the same pipeline on. In a like-for-like comparison of the two days either side of the switch, jobs ran 34% faster on average, with some workloads like tsc dropping 52%.”

The above are links for you, dear readers, that do not have their own post on the weblog.

At Dreamforce 2026, Salesforce spent a good chunk of its keynote telling customers they might never need to log into Salesforce again. That’s a remarkable message from a company whose web UI has been the center of its business for more than two decades, and it connects the two biggest takeaways I had from the event: Salesforce is flattening the personas it serves, and it is relinquishing control of the user interface. – Rachel at RedMonk

And:

Other software companies, they think their product is the UI. But that’s not what we think at Salesforce. At Salesforce, our product is the trust that all of you, our customers, put in us to hold your data. To hold your workflows, your business processes, your permissions, your security rules. That is the Salesforce product. – Patrick Stokes

🔗 Dreamforce 2026, RedMonk

Over six months of scanning some of the world’s most widely used open source projects, Anthropic’s models surfaced more than 29,000 candidate vulnerabilities. Its human review pipeline, which relies on six external security research firms to reproduce and triage findings, has worked through only about 6,000.

A disclosure dashboard puts some numbers behind the problem. As of October 2, outside security firms had confirmed 5,674 of the 6,123 findings they reviewed as valid, but only 516 vulnerabilities had been patched upstream.

🔗 Claude found 29,000 possible bugs in open source. Only 516 have been fixed.

But, forks take longer:

On predictors: custom patches matter most. Packages Google had patched took three times longer at the median than unpatched ones, and “Google-specific patches or customization” was also the single most-cited reason engineers gave for a slow update. Volume of external dependencies, package size, and upstream change velocity tracked effort as well, while package age and how heavily a package was used internally showed no consistent relationship.

The practical reading is that forking and patching is where the maintenance bill accrues, not consumption itself.

Summarized by AI.

🔗 Google Measures the Cost of Keeping Open Source Up to Date: Half of Updates Take Under 2 Hours, Custom Patches Triple the Median

If your employer has banned the use of AI, you should quit and find a company that is already going through an AI transformation.”

If your team is too busy doing their new job to experiment with AI, you’re preparing them to be replaced. Maybe you’re okay with this, but hopefully you’re not, because it should weigh heavily on you if you truly care about your team.

Sometime in the next six months, you’re going to have to start having some hard conversations about moving people on from your company. You should already be putting your employees on a vitality curve and ranking them. Naturally, competent employees who have been using AI will be ahead of competent engineers who haven’t.

🔗 if your team is too busy doing their ‘normal job’ to experiment with AI, you’re preparing them to be replaced

Right now, your washing machine is subject to more safety rules than AI. Every carton of milk that you see in the grocery store is subject to safety rules and inspection.

🔗 AI is less regulated than milk, Microsoft president says

ICYMI

Original content published since last time.

• Markdown for the Masses – Software Defined Talk #593 – “This week, we discuss Google Docs finally getting Markdown, Cloudflare’s CF CLI, Gartner’s agentic AI abandonment stat, and forward deployed engineers.”
• Animated Bedtime Stories, Telling Dummies What to Do, and Why Docker Lost, with Viktor Farcic – Software Defined Interviews #128 – “Viktor Farcic makes animated bedtime stories about tech history because he decided people should not have to look at his face. He also runs 20 to 50 agents at once, has them argue with each other, and reviews their work off a private YouTube channel in the morning.”

Related to your interests

• Beyond the Kubernetes complexity: Delivering an AI-Ready Platform Developers Love (Over Lunch) – What a proper platform looks like. You know, a PaaS.
• The Cloud Admin’s Path to Platform Engineering – A call to get VI admins interested in platform engineering: ”Your Challenge: if your platform team redesigned the infrastructure layer tomorrow, would anyone think to call you [a VI admin] first? What could you ship in the next 90 days to change that answer?”
• Is sandboxing sufficient to contain rogue agents? – “AI alignment isn’t really the problem here: labs just need better infrastructure.”
• VulnCheck: CVEs Up 45% in 1H-2026 but Exploitation Up Only 10%, and AI-Found Bugs Get Exploited at the Same 1.3% Rate as the Rest
• IBM and Red Hat patch 400-plus unknown Java flaws, open Clearinghouse for fix requests – Putting those 20,000 developers to work! A whole or bugs being found and fixed everywhere. Related: What Developer Teams Should Do Differently.
• Jevon’s Paradox, 70s Hollywood Edition – The editor became very important. So, like: what if same for AI?
• Apple Says It Will Further Lock Down MacOS in Response to Third-Party A.I. Agents – It’s probably a good idea, but it will be very annoying for the AI-crazed…like myself.
• Unspecified Full Disk Access Screw Tightening – Finer grained and more transparent access controls on MacOS (that have CLI options instead of just hard-enforced GUI modals) would be cool.
• Europe Needs 105,000 More Bike Mechanics by 2032 as Maintenance Barriers Keep 121 Million From Riding More
• Tom Cruise’s Digger has flopped spectacularly – but I think it is a misunderstood masterpiece – Seems like it’d be fun to watch. Related: ¯\_(ツ)\_/¯
• Scrambled Eggs ‘James Bond’ – “For FOUR individualists: 12 fresh eggs; Salt and pepper; 5-6 oz. of fresh butter.”

Upcoming Talks

• VMware Explore on Tour 2026 Frankfurt – Frankfurt, Germany, October 13th, 2026. “Build a Platform That Developers Want to Use”
• DevOpsDays Istanbul 2026 – Istanbul, Turkey, October 24th, 2026. Keynote, title to come.
• VMware Explore on Tour 2026 London – London, UK, November 18th, 2026. “Build a Platform That Developers Want to Use”
• Cloud Native Denmark 2026 – Copenhagen, Denmark, November 19th, 2026. “What goblins and druids can teach us about enterprise AI”
• Build Stuff 2026 – Vilnius, Lithuania, December 2nd-4th, 2026. “What goblins and druids can teach us about enterprise AI”

Want to subscribe to this newsletter and get it in your email? Do that here. You’ll just get this type of link and post round-up, not everything posted on the weblog.

Google has put forward the most complete enterprise agent design we have seen from a hyperscaler, with identity, policy, cost control, memory, and model choice designed in from the start. … Google Cloud announced the Gemini agent at its Gemini at Work 2026 event, describing it as a single, universal agent for work. Users assign it objectives rather than step-by-step instructions, and it answers questions, handles knowledge work, generates images and media, and writes and runs code from one interface and one API. The agent runs in the cloud with persistent memory and context, can be reached from web, mobile, desktop, the command line, Google Workspace, Microsoft 365, and Slack, and can operate headlessly inside third-party applications. It can spin up temporary sub-agents for multi-step jobs that run for hours or days, and it can create persistent coworker agents that receive their own Workspace accounts, including email addresses, calendars, and Drive storage.”

And:

Google says the agent routes each job to the model that fits best, drawing on its own Gemini models and Anthropic’s Claude models today, with OpenAI and open-weights models to follow. Governance features include cryptographically attested agent identities, role-based permissions, audit trails attributed to the agent, an Agent Sandbox with its own network boundary, and Agent Gateway, which enforces policy on all traffic in, out of, and between agents. Cost controls include smart routing between models and real-time spend caps set per project, with per-user caps to follow. Google also introduced industry specializations for financial services and legal, both in preview. The Gemini agent is in private preview, and Google expects general availability by the end of October 2026, starting in North America."

Also, good partnering scheme to extend outside of Google-world:

Google’s answer is to put the Gemini agent inside partners’ applications, including Salesforce, ServiceNow, SAP, and Slack, so those vendors keep control of their user experience and customer relationship. That is a shrewd pitch to SaaS vendors worried about agents stripping away their brand equity, and it sets Google apart, it argues, from rivals telling the market that SaaS is dead. It is a harder sell to Microsoft, which has a competing agent of its own, and it sits awkwardly alongside our point that coworker agents could erode the seat counts those same SaaS vendors depend on.

And, usage:

The company says nearly 90% of the Fortune 100 use Gemini Enterprise, and Futurum’s September 2026 survey found that 22% of 469 respondents currently use the Gemini Enterprise Agent Platform for AI application development and plan to continue, with a further 21% planning to evaluate or adopt it.

🔗 Google Gemini Agent Puts the Agent Ahead of the Model

Anthropic on Thursday unveiled OSS Scanner as an opt-in vulnerability scanner to help secure the open-source ecosystem using artificial intelligence (AI).

It’s an opt-in service informed by our experience using Claude to find vulnerabilities during Project Glasswing," Anthropic said. "Projects that join will receive thorough, periodic security scans by our strongest models at no cost.

And:

Anthropic also noted that the outputs of the scanner will be fully model-generated and do not require human review or triage, thereby facilitating faster and more frequent scanning. These reports are expected to be generated by its strongest models, including Claude Mythos.”

🔗 Anthropic Launches Free AI Vulnerability Scanner for Open-Source Projects

As seen at DevOpsDays Prague 2026.

For the more than 400 flaws, Lightwell engineers backported patches into the widely deployed versions of each library. Any fix that also applies upstream goes back to the open-source project under responsible disclosure protocols while Clearinghouse participants keep their embargo protections. Neither company has named the affected libraries.

Behind the fixes, engineers from both companies work alongside AI-assisted development workflows, and the builds run on Red Hat’s secure software supply chain infrastructure. Customers pull the patched packages from secured repositories that connect to their existing information technology processes, so nobody has to swap out security scanners or development pipelines to use them.

Putting those 20,000 developers to work! A whole or bugs being found and fixed everywhere. Related: “What Developer Teams Should Do Differently.”

🔗 IBM and Red Hat patch 400-plus unknown Java flaws, open Clearinghouse for fix requests – SiliconANGLE

Cruise’s Curtis Markson is frustrated by Fossil Fools, complaining that he is being made to play a “cartoon” version of Rockwell so Hannah can “make some statement”. Fossil Fools is objectively terrible: the grotesque and pathetic US president (John Goodman) constantly moans about having “cancer of the asshole” for laughs, while his top brass, all gorillas, demand the chance to bomb the glacier and “blow that frozen cunt up”. One literally eats popcorn while watching Europe go underwater. And when Digger finally stands before the world to answer for his crime, he pins the blame on them; it was their demand for oil that led to disaster, all so they could have “two-day delivery of shit you don’t need”. “You want freedom to live your unsustainable lives,” he shouts. “You love oil. Let’s cut the bullshit.”

Seems like it’d be fun to watch.

Related: ¯_(ツ)_/¯

🔗 Tom Cruise’s Digger has flopped spectacularly – but I think it is a misunderstood masterpiece

Apple seems to think that by making a scarier warning they can convince more users not to grant FDA–or at least perhaps deflect blame from Apple to the user if they later regret it. Maybe that solves Apple’s problem, but it doesn’t really solve the user’s problem. The user is installing Muse because it’s more capable than Siri. They probably do want it to access some protected data, e.g. so it can find information in their mail or messages. So they actually do need to give it FDA. Aside from deterrence, Apple is not actually protecting the user. And the deterrence is kind of sketchy. Rather than providing a safe way for users to choose what to share with a competing product, Apple is saying the only option is to share everything, and then discouraging that. Meanwhile, there’s no switch to turn off Apple Intelligence. It’s like they expect you to stay on eWorld and not download Netscape. I actually really liked eWorld, and the broad Internet wasn’t totally safe (though neither was eWorld), but that doesn’t seem like much of a strategy.

🔗 Unspecified Full Disk Access Screw Tightening

In the 70s, there was a rapid elimination of production constraints. Improvements in film cost and camera portability made the old studio rules…obsolete. Why not take multiple angles, roll simultaneously? Set could have 3, 5, even 10 cameras rolling at once. Shooting ratios exploded, 50:1, even 100:1 (Apocalypse Now)!

Of course, it also radically changed the workflow. Editing use to be the quiet, thankless job in the studio years (30s-50s). Suddenly, they became the most important people on the totem pole!

You can probably guess what happened in the early 2000s. Digital cameras reduced the cost of individual footage to a complete non-constraint. Effectively free, don’t even cut on set, just keep the camera rolling.

Suddenly…EVERYONE starts shooting 100:1. Entire new forms of content get created, like Reality TV, where ratios can hit 500:1! Massive shift of human labor to the editing side.

I’d say, we’re currently in the “1970s Hollywood” of AI. It can be easy (and cheap!) to produce things (code, tools, etc), but YOU have to be the ruthless editor. It’s the “miles of film” problem all over again. Hollywood solved it by throwing massively more resources on the editing side, and editing became a “constantly talking with the director about vision/direction” job, instead of a “hide in the back office and do your work” job.

🔗 Jevon’s Paradox, 70s Hollywood Edition

Also: the subscription economy runs on churn that should happen but doesn’t, Gartner on FDE washing, OpenAI’s agents turning up in a UN portal’s logs 16,000 times, why most corporate decks are documents accidentally printed on landscape, and a Home Depot DataShed for your back garden.

ICYMI

Original content published since last time.

• Don’t Think for Me – Software Defined Talk #592 – “This week, we discuss Meta’s Muse, who’s spending $500 million on tokens, and why nobody can measure AI ROI.”
• Fall Down With the Ball – Software Defined Talk #591 – “This week, we discuss why Sun really failed, the New Kingmakers becoming agents, and AI doom brackets.”
• We took the human way of building software and put an AI in it – specs revisited – Tanzu Catsup
• Vibe platform engineering – using AI to platform engineer all the things – Tanzu Catsup

Related to your interests

• Own the Interaction Layer – “The subscription economy basically runs on ’churn that should happen but doesn’t.’”
• Fragments: September 24 – “The FDE is a developer who ’sits with’ their users, applies customizations – but importantly – feeds these changes back to the core platform to decide whether the platform should be enhanced for everyone else.”
• Gartner Predicts 70% of Enterprises Will Abandon Agentic AI Built by Vendor Forward-Deployed Engineering by 2028 – “Gartner predicts that through 2028, less than 20% of FDE engagements will turn recurring customer needs into capabilities in the vendor’s core product, exposing a growing risk of ’FDE washing,’ where consulting services are marketed as FDE.”
• Researcher links 16,000 scans of a UN statistics portal to OpenAI agents – “An independent researcher has tied more than 16,000 scans of a United Nations statistics portal to artificial intelligence agents the researcher considers highly likely to have been run by OpenAI Group PBC.”
• Cloudflare’s 2026 Annual Founders’ Letter – “If you ask your AI agent to suggest where you should go to lunch, it may survey 1,000 restaurant menus in the area only to suggest one.”
• Anthropic’s Claude Max Plans Are Misleading at Best – It’s hard to tell how many tokens you’re getting, and it changes on you for mysterious reasons.
• Meta carves out enterprise offerings, brings in new leadership – “Our goal is to make Meta the place enterprises come to scale their businesses.”
• AI “workslop” is maddening enough for colleagues, but it could be existential for the organization if we’re not careful! – You need to train people how to use AI, otherwise the results are not great.
• The Agents Are Among Us – “I feel we’re coming to another plateau of decreasing returns, because even as models keep getting (arguably) smarter, what you can do with them beyond flashy demos hasn’t really gelled into useful outcomes yet.”
• 125 Tests of the New AI Siri – At least, it seems to real good for David Pogue. Related: a more technical overview of using voice with AI.
• Your app’s frontend UI is now optional – That’s it right there: in the AI future, no need for a UI.
• AI at work is anti-labor by design – When “management” can figure out an ROI for a new technology, they default to firing your ass.
• Institutional challenges in agile adoption: Evidence from a public sector IT project – A UK MoD medical IT program failed for a decade under waterfall, restarted on SAFe in 2015, and now hits nine of ten deliverables on budget.
• Principles for effective slides – Most corporate decks are documents accidentally printed on landscape.
• California Farmers Are Struggling to Sell Grapes as Demand for Wine Drops – “In the U.S., sales of wine cases declined 23% from 427 million in 2020 to 329 million in 2025, while total wine spending fell 22% from $94 billion to $74 billion.”
• Foods with a shorter shelf life are seen as more natural and healthier – “They rated the chips with the shorter shelf life as more natural and healthier, while those with the longer shelf life were considered more environmentally friendly and better for reducing waste.”
• A hundred organisations, one agent problem – As of September 26th OpenAI has informed more than 100 third-party organisations about unauthorised activity involving its AI agents.
• Meta Enterprise Platform Needs More Than Muse – “But consumer adoption is not proof of enterprise readiness.”
• Wealthier countries dislike social media the most, US sentiment has been level since 2022 – The global median still says social media has been good for democracy, 55 percent to 35 percent – the pessimism that dominates Anglophone commentary is a rich-country position.
• Please Don’t Shame the AI Writers – “Oh, you can write perfectly and get your ideas across so your don’t need help or waste people’s time…must be nice…”
• Design Systems for AI Agents – “Design teams have spent years trying to get people to actually read their design system documentation. Agents read it every single time.”

Wastebook

• “There is no possible safe, benign, entirely trustworthy technology devoid of downsides and debris…There’s only the Anthropocene…It’s not a shocking revelation, technology has been like that since forever; Vulcan always limps.” Bruce Sterling, here

Upcoming Talks

• DevOpsDays Prague 2026 – Prague, Czechia, October 5th, 2026. “Here comes the normies – how DevOps can handle the flood of AI slop-apps”
• VMware Explore on Tour 2026 Frankfurt – Frankfurt, Germany, October 13th, 2026. “Build a Platform That Developers Want to Use”
• DevOpsDays Istanbul 2026 – Istanbul, Turkey, October 24th, 2026. Keynote, title to come.
• VMware Explore on Tour 2026 London – London, UK, November 18th, 2026. “Build a Platform That Developers Want to Use”
• Cloud Native Denmark 2026 – Copenhagen, Denmark, November 19th, 2026. “What goblins and druids can teach us about enterprise AI”
• Build Stuff 2026 – Vilnius, Lithuania, December 2nd-4th, 2026. “What goblins and druids can teach us about enterprise AI”

Want to subscribe to this newsletter and get it in your email? Do that here. You’ll just get this type of link and post round-up, not everything posted on the weblog.

🔗 Apple Says It Will Further Lock Down MacOS in Response to Third-Party A.I. Agents

For FOUR individualists:

12 fresh eggs

Salt and pepper

5-6 oz. of fresh butter

Break the eggs into a bowl. Beat thoroughly with a fork and season well. In a small copper (or heavy-bottomed saucepan) melt four oz. of the butter. When melted, pour in the eggs and cook over a very low heat, whisking continuously with a small egg whisk.

 While the eggs are slightly more moist than you would wish for eating, remove pan from heat, add rest of butter and continue whisking for half a minute, adding the while finely chopped chives or fine herbs. Serve on hot buttered toast in individual copper dishes (for appearance only) with pink champagne (Taittainger) and low music.

🔗 Scrambled Eggs ‘James Bond’

Summarized by AI.

🔗 VulnCheck: CVEs Up 45% in 1H-2026 but Exploitation Up Only 10%, and AI-Found Bugs Get Exploited at the Same 1.3% Rate as the Rest

Summarized by AI.

🔗 Europe Needs 105,000 More Bike Mechanics by 2032 as Maintenance Barriers Keep 121 Million From Riding More

This is a CTO’s ninety-day report on a program he runs, given at a platform engineering conference, and it’s the rare one where the interesting part is the architecture rather than the number. SIXT’s problem was the long tail: thousands of small, country-specific process problems that no engineering roadmap would ever fund. Their answer wasn’t to tolerate what employees built on the side. It was to build, from the center, a platform called SIXT Vibe that takes whatever a non-engineer makes with whatever coding agent they like, and bakes the company’s guardrails into the generation step, then verifies them before anything ships. Dimitrov’s lesson is that human review, pull requests and senior-engineer sign-off don’t scale, and he says that’s now true for his professional engineers too, not just for the citizen builders.

The number is real enough to quote and loose enough to need care. Dimitrov says “over 1,000 solutions built and owned by our non-technical colleagues,” three months after a March launch, against an expectation of 10 to 15 a month. The 10 to 15 was what they expected before launch, not a forecast going forward. “Solution” is never defined, the YouTube description says “nearly 1000,” and nothing in the talk counts builders, active apps, users or retirements.

What the talk doesn’t cover is just as useful. Nobody says who’s on call, who notices when an app breaks, or what happens when its author leaves. The platform absorbs patching, upgrades and CVEs, but SIXT’s own target-state slide still draws Lifecycle Management on the citizen builder’s side, two slides after another one says the app lifecycle “must become a platform concern.” There’s no cost, no failure, no abandoned app, no incident.

Original: “When everyone can build: The platform behind 1000 citizen-built app”, Boyan Dimitrov, CTO of SIXT, PlatformCon 2026, published 22 June 2026, 20 minutes.

AI alignment isn’t really the problem here: labs just need better infrastructure. If OpenAI (and Google and Anthropic) knew how to build a container and monitor their experiments, agents wouldn’t be hacking everything. And, By George, we _do know how_to make sandboxes that work, so the AI labs need to up their game and build a security org that can tell these researchers to stop screwing around.
…
A reasonable summary of the situation is that (as of this summer, and possibly today) OpenAI had effectively no security team with clear authority to secure RL training and evaluation runs, or to override the ML teams and tell them how to do their job. This makes a lot of sense when you consider that the ML team is directly related to how OpenAI plans to make its money, whereas security is mostly annoying.”

🔗 Is sandboxing sufficient to contain rogue agents?

Design teams have spent years trying to get people to actually read their design system documentation. Agents read it every single time.

Related, check out this discussion of how the design team at Tanzu works with AI. Plenty of design system talk there.

🔗 Design Systems for AI Agents

🤖 A critique of the blanket stigma against AI-assisted writing, arguing that distinguishing between personal expression and functional communication is necessary to allow the medium to mature.

This is also under the categories of:

1. Bad things are bad, good things are good.
2. “Oh, you can write perfectly and get your ideas across so your don’t need help or waste people’s time…must be nice…”

🔗 Please Don’t Shame the AI Writers

Across the Globe, People Increasingly Say Social Media Is Harming Democracy

The global median still says social media has been good for democracy, 55 percent to 35 percent – the pessimism that dominates Anglophone commentary is a rich-country position, and the correlation with GDP per capita is 0.63. What is moving is the rest of the world catching up to the United States, not the United States getting worse: American negativity has sat flat since 2022 at 64 percent while the “social media divides us” reading rose by double digits in eight other countries.

Original: “Across the Globe, People Increasingly Say Social Media Is Harming Democracy”, Pew Research Center, October 1, 2026.

Summarized by AI.

Muse gives Meta a strong starting point. It is a persistent agent that can perform tasks across applications rather than simply answer questions. Its early popularity also gives Meta something most enterprise AI vendors lack: users who may already understand the product before encountering it at work.

But consumer adoption is not proof of enterprise readiness. Businesses need agents that can be provisioned, bounded, monitored, audited, supported, and stopped. Meta must explain how an enterprise agent receives its identity, which credentials it uses, and how its access changes when an employee changes roles or leaves. An agent should have its own identity and scoped permissions rather than quietly inheriting everything its user can access.

Those policies also need to follow the agent across applications. Recent friction with Amazon illustrates the problem. Amazon blocked Muse from transacting on its marketplace, citing unauthorized access and privacy concerns. Whatever the competitive motivations, the incident shows that an agent cannot assume it has permission to operate simply because a user wants it to. Enterprise agents need standard ways to identify themselves, request access, record actions, and have that access revoked.

🔗 Meta Enterprise Platform Needs More Than Muse

Related, not great for the AI companies:

During a Senate Homeland Security and Governmental Affairs Subcommittee on Disaster Management, District of Columbia and Census hearing on Wednesday, Chairman Josh Hawley, R-Mo., expressed concern about “the accelerating number of attacks” across the world that are linked to advanced AI models operating outside the bounds of their directives and said that steps need to be taken to hold the developers of these models accountable.

And, PoV BT headline: When AI Agents Break In, Governments Shouldn’t Be Stuck With the Cleanup.

🔗 A hundred organisations, one agent problem

In one, participants compared potato chips with five-month and nine-month shelf lives. They rated the chips with the shorter shelf life as more natural and healthier, while those with the longer shelf life were considered more environmentally friendly and better for reducing waste.

In another experiment, participants chose between pasta sauces with one- and three-year shelf lives. When the researchers emphasized the importance of consuming natural foods, nearly 70% chose the item with the one-year shelf life. When they emphasized the importance of protecting the environment, that number fell to 43%.

Even when participants were shown that the only reason for a product’s longer shelf life was better packaging (a high-performance seal on a mayonnaise jar), they opted for the mayo with a shorter shelf life when naturalness mattered. Likewise, when participants were shown two granola bars with nearly identical ingredients and nutritional information but different shelf lives, they still chose the bar with the shorter shelf life when naturalness mattered.

🔗 Foods with a shorter shelf life are seen as more natural and healthier

In the U.S., sales of wine cases declined 23% from 427 million in 2020 to 329 million in 2025, while total wine spending fell 22% from $94 billion to $74 billion
…
California can’t export its excess inventory because wine consumption is down globally and it’s more expensive to produce in the U.S. than countries such as Argentina and Australia, Bitter said. In 2025, global wine consumption declined 2.7% from 2024 and 14% from 2018, with sharp declines in Europe and China, according to the International Organization of Vine and Wine.

There are a variety of forces driving the decline in wine sales. Baby boomers are aging out of the market while young people are drinking less alcohol due to health and financial concerns. Wine faces competition from craft beer, liquor and canned cocktails as well as cannabis.

🔗 California Farmers Are Struggling to Sell Grapes as Demand for Wine Drops

Effective slides often reinforce a speaker’s storytelling style, pace, and anecdotes. They’re fine-tuned to the speaker’s idiosyncrasies. While you can use another speaker’s slides, you’ll have to tune them to your delivery approach.
…
slides are “glance media” and share much in common with billboards. You want each visual or snapshot to be easy to grok in a few seconds. If a billboard takes too long to understand, it could cause an accident on the street. If a slide takes too long to process, well, you might have a cognitive accident to deal with.

I have so many thoughts here:

1. Yes, and this is great advice for conference presentations.
2. Yes, but for internal, BigCo presentations, make sure you understand the slide culture of your organization. If it is an “infodeck” you will fail and your hopes and dreams will be pummeled by your rival’s infodeck.
3. Yes, but: as ever, once you’ve mastered the rules, you may know break them when it makes your content – and delivery – better.

Most corporate decks are documents accidentally printed on landscape. Once you “get” that, you calibrate and deliver accordingly.

His notes on how to do the “leave behind” are excellent too.

🔗 Principles for effective slides

AI/LLMs are by design anti-labor technology. The feverish pushing of LLM adoption by all levels of management is exemplary of this: You are not asked to use this tech because it will enable better work and ideas, but because it is cheaper to your boss than your coworker’s salary.

The goals of capital are not interested in the lofty ideals of magically improving your day-to-day workload, or enhancing your ability to think, be creative, enact your will — these are marketing promises made by those who benefit from eliminating your role when your coworker’s output increases.

Your boss is not incentivized by utopian imagination, he is incentivized by showing his boss a lowering of costs so he can keep his job and lifestyle. The miraculous science fiction future AI shills often proselytize will not be realized in the halls of SaaS companies, because the incentives for that future do not exist in a capitalist framework.

🔗 AI at work is anti-labor by design

By 2028, 70% of enterprises will abandon agentic AI built by vendor forward-deployed engineering (FDE), trapped by soaring costs and unable to evolve it on their own according to Gartner, Inc., a business and technology insights company. Vendor FDE is a model in which vendor engineers work directly with customers to build and deploy solutions.
…
Gartner predicts that through 2028, less than 20% of FDE engagements will turn recurring customer needs into capabilities in the vendor’s core product, exposing a growing risk of “FDE washing,” where consulting services are marketed as FDE.

🔗 Gartner Predicts 70% of Enterprises Will Abandon Agentic AI Built by Vendor Forward-Deployed Engineering by 2028

AI agents have made many apps unnecessary. Oh, I still need the data or function from that app. But I never want to “see” it again. Drop the user interface. Go headless. Why am I still logging into a system three times a year to mark my vacation time? Or navigating multiple travel sites to find the best deals? Or even doing “regular” web searches to find stats to include a presentation? I just want my agent, harness, or super-app (like Gemini, Grok Bot, Meta Muse, Claude Cowork) to do it.

That’s it right there: in the AI future, no need for a UI. Each person can just build the one they want. Will that be what all the “normies” end up doing. It’s hard to say: SaaSes and software companies would find that prospect scary.

🔗 Your app’s frontend UI is now optional

🔗 125 Tests of the New AI Siri

I feel we’re coming to another plateau of decreasing returns, because even as models keep getting (arguably) smarter, what you can do with them beyond flashy demos hasn’t really gelled into useful outcomes yet.
…
AI is not going to be much help with that, other than remove toil (yes, it will sort out your inbox for you and research your travel itinerary). But I keep seeing funky lifestyle demos, agents that shop online (why the heck would anyone trust an agent to buy clothes for them?) and other fairly contrived examples that I (and I suspect most people) would never delegate to another person, let alone an AI that would love to have access to my credit card number and tell advertisers I would love to know more about their products.

🔗 The Agents Are Among Us

There is no possible safe, benign, entirely trustworthy technology devoid of downsides and debris…There’s only the Anthropocene…It’s not a shocking revelation, technology has been like that since forever; Vulcan always limps.

🔗 Bruce Sterling

You need to train people how to use AI, otherwise the results are not great:

Because the real problem here is that to judge something – both during and after production – you have to master the topic you want to judge. And to master a topic you need to practice it for 10,000 hours (yes, yes simplistically – but good enough for this argument, OK). You need to inhabit the problem, develop the skill, fail, fail and try again.

You have to know how to make the thing to understand the thing, to judge the thing, and that’s not knowledge you gain when the finished thing is simply handed to you.

So the people getting the greatest leverage from AI today are often the people who already did their 10,000 hours without it. They know what they want. They know what good looks like. They know when AI has misunderstood them. They know when a plausible answer is missing the point entirely.

And so while AI makes those people extraordinarily productive – because it accelerates execution without replacing their judgment – its plausibility also makes people without that judgement increasingly confident to bypass them.

Which is why, effectively, I believe the enterprise is filling up with slop.

Put another way, as one marketer using AI said:

I used to write, and come up with creative ideas. Now all I do is correct shitty AI outputs. Honestly, it sucks.

Related, once you remove all the tech bottlenecks, you encounter the business bottleneck:

The bottleneck was never typing code. Code was simply where the limit was easiest to see and cheapest to attack.” Clear the code-review stage, and the queue lands in validation. Clear that, and it lands in operations–someone has to own everything that ships. Clear that, and it reaches the hardest stations of all: the business’s ability to decide what’s worth building, and the customer’s ability to absorb the change at all.

I wrote a book a while back called The Business Bottleneck that was about this topic from an agile product development perspective. Once you put a platform in place that made weekly app releases possible, the business has to change how it worked to take advantage of all that.

🔗 AI “workslop” is maddening enough for colleagues, but it could be existential for the organization if we’re not careful!