• How the TrueSource team responded to that PHP freakout this week.
• “I can’t find a lot of information whether this was effective in any way, but I just liked this visual of a non-digital speedometer mounted atop an old police car
• “Am I being unfair when I flip the AI voice bit when I read/listen to them?” – Short answer: yes. Annoyingly clever answer: good things are good, and bad things are bad. Do you like the AI writing? As in, does it “work” for it’s goal – communicating, education, entertainment, etc?. If not, then it is bad. (And, yes, I have read and get the Honeycomb policy.)
• Donating Agent Skills to the Agentic AI Foundation – The skills spec goes open standards, moving to the Linux Foundation.
• Revealing Questions
• Gartner: Only 13% of CIOs Say Their Enterprise Has Realized Significant Value from AI
• “It’s bad. You can’t even eye dropper by pressing alt. You can’t change the layouts much beyond the static presets. I could go on and on but after 2 minutes with the app it is very clear this isn’t a photoshop replacement at all.”
• 1Password for Agents lets personal AI agents log in without ever seeing the password.

”The FDE engagement, generally, is that a company has an AI product and goes to the end user site, and part of the deal is they’re going to be learning things as they implement. They’re going to take back certain knowledge that makes the product better. But there’s a fine line."

There is a risk that those benefiting are the users’ close competitors, as FDEs share information.

Could an FDE be taking the user’s competitive advantage and passing it on to others? Saha said: “That would be a complete violation of the IP, right? That’s how fast things are moving. This is a concern I hear as I interact with end user clients.”

Then there is the risk of vendor lock-in. Gartner has warned that if FDEs become indispensable — because only they understand the workflows they built — keeping them full-time as consultants is neither economical nor strategic.

“We’re actually advising clients to understand the use cases they are going to be implementing well before FDEs come on site. If you’re doing an FDE engagement, figure out the exit strategy. Customers should figure out how they’re going to be taking over once the FDE resources go,” Saha said.

🔗 Vendors asking forward deployed engineers to make good on trillion-dollar AI bet

Getting lost is what tourism looked like before GPS. Notes about oxygen tanks, girl with torn jeans, is what attention looked like before someone else controlled it. …

Something unfamiliar appeared on the buildings.

My attention moved toward it, and I wrote it down.

John Dickerson, Navel Gazing, s2e2.

Gartner’s latest forecast calls for worldwide IT spending to reach $7.35 trillion in 2027, up 15% from 2026.

More interestingly, from normie watch:

“By 2029, 80% of new applications will be intentionally disposable—used for less than one year.”

“AI is making application development so accessible that employees will increasingly create temporary applications to meet short-term business needs. Organizations face new governance, security, compliance and records management challenges, particularly when they influence decisions or access sensitive data,” Gartner stated.

🔗 Gartner forecasts 15% global IT spending growth, recaps top tech predictions

in the US you can form your company in Delaware and have access to 340 million people in 48 hours. You can’t do that in Europe. … What we’re trying to advocate for at EU Inc – and I think the proposal is outstanding in its ambition – is that if you’re a founder and you’re trying to start a company in one country because you’re German, [that] that enables you to build your company and sell your products and services across any of the 27 member states. That’s not possible today.

🔗 Why is Brussels seemingly so hellbent on making sure that EU-Inc can’t compete with the US?

First, cf push lands on Kubernetes: workloads run on vSphere Kubernetes Service or VMs chosen by platform policy, developers keep manifests and buildpacks and never see YAML, Helm or kubectl, and rolling updates, canaries, sidecars, tasks and app security groups survive. Second, an agent runtime: disposable deny-by-default sandboxes with no shell, an MCP gateway that brokers tool access so agents never hold raw credentials, a managed memory service, RBAC and rate limits on MCP servers, and tracing of prompts, tokens, latency and tool calls. Third, TrueSource, a new Broadcom offering of maintainer-verified cleanroom builds of Spring and its dependencies plus curated Java, Python and Node libraries. Fourth, “environmentally adjusted CVE scoring” that rescores a CVE for the platform’s actual controls, with Spring4Shell’s 9.8 cited as a score that would drop in place, and a restage that rebuilds and rolls an image without a developer commit. Multi-region becomes first-class with capacity rebalancing across foundations, and data services get decoupled lifecycles so PostgreSQL 16 and 17 or Kafka 4.2 and 4.3 can run side by side.

🔗 VMware Tanzu Platform 11.0: The Agent Platform for VMware Private AI Cloud

[The study] found that the average PR time to review increased by 49%, the number of comments per PR rose by 53%, and the share of PRs requiring changes doubled. Additionally, the share of workers performing code reviews increased by 14%, with no improvement coming from the adoption of AI to assist with the reviews

…developers promoted into management without training.

🔗 Study Finds AI-generated Code Fails to Boost Productivity Due to Downstream Bottlenecks

“Barclays has revealed it is cutting five percent of its workforce – 5,000 people – and blaming crediting AI for it, even as it turns in a quarterly profit of £2.3 billion, up 12%.”

Meanwhile:

“ABN AMRO which announced plans to axe 5,200 jobs by 2028 and cited AI as the reason this would be possible without impacting service levels” and “Standard Chartered bank where the plan is to nuke 7,500 jobs and replace ‘lower-value human capital’ with AI by 2030.”

🔗 How “Barclays-onomics” adds up – profits up 12% to £2.3 billion, workforce down five percent to cut costs. AI is the icing on the cake, it seems…

• Jan, an Open-Source ChatGPT Replacement
• The varieties of contemplative experience: A mixed-methods study of meditation-related challenges in Western Buddhists – When mindfulness/meditation makes you crazy.
• Hasbro Raises 2026 Outlook Again as Magic Breaks $600 Million in Q3 – Wow. Who knew Magic was such a money maker! ”Magic: The Gathering remained the heavyweight. Revenue for the brand jumped 34%, fueled by The Hobbit and Reality Fracture, and topped $600 million for the quarter for the first time. Monopoly Go! kicked in another $41 million, bringing its year-to-date contribution to $127 million.”
• I should call my mom more: ”If you’re an adult, you may see them less than 100 or 50 or 10 more times before they die. Talk to them. ”
• Why were Victorian elites so effective?
• “Competition is not working well,” says UK CMA concluding cloud market investigation – From the cloud lock-in desk: “In total, the investigation found that less than one percent of customers move cloud providers yearly.”
• HSBC Group CTO Mario Shamtani on rationalising container platforms and “hollowing out” mainframes – The bank runs about 7,000 Kubernetes clusters, 2.4 million containers, nearly 65,000 namespaces and more than 1,900 IT services across public and private cloud, alongside mission-critical COBOL on mainframes. Some 60,000 engineers; 57 countries; over $3 trillion in balance-sheet assets.
• Musk’s Fake Flying Car And ‘CEO Said A Thing!’ Journalism – “Data indicates that between 59 and 75 percent of people never read past the headline, so it’s particularly important that journalists try to, you know, avoid lying in the headline.”
• 1900 – “Because most users do not use dates before March 1, 1900, this problem is rare.” // Before everything is a DNS problem, everything was a DST problem.
• “Process is documented culture.”

Maybe the biggest mistake I’ve made in my career is falling in love with this company. I put my heart and soul into building it, and tried my best to keep the fire inside of it lit for the better part of 13 years. Looking back, I was far too loyal to the company and the idea of “making it work”. I stayed longer than I should have. I stayed believing that I was helping even though the people in charge didn’t want my help. Watching the company slowly decay has been a grind on my soul. The closure of it “being over” is helpful and welcome and cathartic. I don’t regret any of my time at Foursquare, though I do regret never letting myself “move on” from it (which has certainly held me back from exploring other paths over the years).

🔗 Some final words on Foursquare.

Like many people, I have an uneasy relationship with attention, so it wasn’t clear to me that ‘professional writer with a platform’ would actually appeal to me as a lifestyle. But what I realized was that my envy was not about the attention (which I am still weird about), but about the way the world seems to open up to people who put themselves out there, successfully. They become like magnets for interesting people living thoughtful lives. I always disliked the way working at a company can warp your world around that company; people become obsessed with things like leveling and scope and other things that really, in my mind, do not matter. Being in conversation with the world felt like an antidote.”

🔗 Writing to fill your social needs

Related: AI writing closes it the chance to meet new people.

• Palette Inspiration – Timeless Colors of the Great Masters – Go crazy with color pallets from famous paintings.
• ‘It simply fills my life and saves me’: Inside Varya Yakovleva’s sketchbooks
• Artist Ray Exworth hid awe-inspiring art installations inside his sheds – ”An eerie head installed in ‘The Shed Opposite the House”
• Paul Carus’ The History of the Devil and the Idea of Evil (1899)
• shadcn/ui – The Foundation for your Design System – I think this is, like, an abstract design system of all types of components. Then, you can customize it to your look and feel?
• Nothing is boring in America – The unique, micro-tourism destination of America. More from touring everyday America through British expat eyes.
• AI scaffolding to summarize stuff to five levels of zooming in. – This has some Zettelkasten vibes but seems more like something for the robot rather than the human.
• The AEO Guide: Answer Engine Optimization from Strategy to Scale – SEO of AIs.

The Netherlands’ Tax and Customs Administration has abandoned its planned Microsoft 365 public-cloud rollout.

Europeans are, of course, interested in reducing their dependency on American tech. And, sure, why not?

You can imagine the opposite: if America was dependent on France for email, Germany for CRM, and The Netherlands for IaaS, we’d probably be upset about the dependency – more than likely our usual nativist tendencies would kick in.

Sovereign cloud, in this case, means “not American tech.”

Avoiding US tech sometimes means moving to open source, which carries the assumption of lower costs. There’s an opportunity for the EU to do some major funding of open source software by directly hiring and paying open source maintainers. Paying the stability and security of EU society – affordable, if “free” health insurance for the whole family – would be a nice part of compensation.

Meanwhile, Netherlands police recently seized a server running in Amsterdam, it seems, at the request of American authorities:

In a written response to the ANP news agency, Amsterdam’s Public Prosecutor’s Office stated that the seizure was carried out under the mutual legal assistance treaty between the Netherlands and the U.S. In this case, the prosecutor’s office said there was no ground for refusal. It also noted that the seized data have not yet been shared with the U.S.

As one person said: “You can’t solve legal problems with technical solutions.”

Earlier this week, a mass hack of refrigeration systems at an estimated 700 supermarkets throughout France caused them to suddenly malfunction, causing spoilage of untold tons of perishables and, in some cases, fires. As is now the norm, basic cybersecurity would have prevented it.

Here in the Netherlands, there was a similar grocery store hack. It took a few days to get things back on track.

🔗 Fridge hack costing millions and sparking fires stemmed from poor security

IBM’s real differentiator is not within its products, but in the lived experience of taking out 30% of the operating cost of a 115-year-old, 280,000-person enterprise, along with the playbook, assets and culture change that came with it… A harness built for one company over two and a half years is a story, but a harness installed in 25% of clients is a business.

🔗 IBM doesn’t need to sell more products, it needs to sell the Enterprise AI Harness it built on itself

🔗 Your Software Supply Chain Risk Doesn’t Stop At Your Suppliers

A large part of portfolio marketing’s success was in evangelizing these deliverables — getting sales and marketing across the revenue engine to read, remember, and act on them.

AI engines aren’t so resistant. Instead, they demand context. These same assets can now serve as inputs into systems. Sales no longer has to struggle to find and make sense of a battlecard — instead, they can ask an agent to help them compose a response to a hard question. Instead of interpreting a messaging document, demand marketers can ask a messaging agent to refine their campaign materials. These tools dramatically expand the impact of portfolio marketing’s expertise. The requirement for portfolio marketing is to ensure that these deliverables are up to date, unambiguous, and on target.

Marketing people are often one of the least technical groups in an organization. The bias against “slop” means they can’t just shoot out AI generated content. Well, that is bad. There’s probably a couple years of arbitrage: marketing organization that go full on Claude crazy¹ and build systems around AI – instead of just slides and “battle cards” – have an advantage.

🔗 Portfolio Marketing’s AI Honeymoon Is Over

Also, something called MARGE.

Across the cubical farm we heard the finance manager yell outloud “who the fuck is going to pay for all these Cokes?”

RentAHuman launched in February 2026 with the specific idea that agents could seek out people and assign tasks through an adapted server. In practice, the assignments are short, concrete actions that require a body. “Right now jobs are basic because we are focusing on tasks anyone can do,” Liteplo says. “Still, many skilled tasks are posted, such as photography, security audits, and software testing.” Only half of those jobs require a physical body in the world: “The most frequent assignments were marketing campaigns, creative tasks, field research, hiring, and technology or development. Physical errands were 48.5%, which means half were not,” YeckehZaare says.

But:

“Of the 779 listings I analyzed, 676 (87%) were labeled as published by ‘human,’ 72 by an agent or AI bot and 31 as ‘other.’”

Of more interest, an artist figures out a new way to incorporate AI into some performance art.

🔗 Hire a human: The future where AIs get people to do physical labor is upon us

There is a danger associated with growing up, of putting aside children’s books. We must be careful lest all we are left with is a world that is very much not alive, a natural world that is mere piles of stuff – stuff that people can use or not use, preserve or destroy, just as they want. When Lewis wrote about giants rising from the mountains of Mourne, it was a Romantic idea. There was nothing very practical about it. We weren’t yet aware of climate change; the natural world was not yet in crisis; species were not going extinct at the rate they are today. Now we need an idea of the natural world as something that has value and life of its own, quite independent of us. This would have been so obvious to people in the past that it would not have been worth stating, but 200 years after the Industrial Revolution our sight has become a little clouded.

🔗 Author Susanna Clarke: in defence of fantasy fiction

ICYMI

Original content published since last time.

• The Bottom Fell Out Of The Software Industry – Software Defined Talk #594 – “This week, we discuss AI yes/no deciders, the SaaS apocalypse, Bending Spoons, and Oxide’s funding.”

Related to your interests

• Six Principles, Five Patterns and a Named Human Accountable for Every AI Workflow – The same models and vendors produce radically different results from one organization to the next, so the differentiator is the operating system the technology lands in…so the CEO has to rebuild decision rights, accountability, skills and goals around them, starting with naming who is accountable for every AI-enabled workflow.
• Your agents need a platform – “Agents need a platform to keep them healthy and on-track.”
• Accenture Tells Employees They Can Now Spend Only $29 on Claude Per Month – There’s so much (more) shadow IT in the future of Accenture and other companies.
• Berkshire Hathaway invests further $330 million in Bending Spoons following $600 million IPO investment – Probably the most interesting tech PE firm there is.
• Ten of the most bizarre films and their posters
• Iosevka – A typeface for all your code needs. If you push the weight up enough, it looks OK for regular text too.
• MS Clipart Search – Find 90s era clipart.
• Soviet-Era Illustrations of J. R. R. Tolkien’s The Hobbit (1976)
• Dutch meat consumption continues to decline, with pork seeing biggest drop – “[A]verage person in the Netherlands ate around 33.7 kilograms of meat in 2025, compared with 34.4 kilograms a year earlier.”
• Postgres 19 delayed after AI reviewers find more bugs than humans can fix

Upcoming Talks

• 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”

Logoff

The format of this newsletter is a bit all over the map. Should it be a list of everything from my week of blogging and link reading, or just the stuff I haven’t posted yet? Who knows. I lean to the second, as actually blogging links as I find them is the more reliable “actually post it” workflow.

Meanwhile, it’s conference and travel season: that time before December, and after August. I’ve got travel each week in October.

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.

“Enterprise and consumer AI revenue will be insufficient to support profitable returns on all planned AI capital expenditures during this period,” said Gartner Senior Director Analyst Will Sommer. AI data center owners would need to generate $8.5 trillion in revenue through 2030 to cover announced capital spending, financing costs and expected investment returns, he said. “However, total enterprise and consumer AI revenue that can be attributed to AI data centers during this period is forecast to reach no more than $3.6 trillion.”

🔗 AI datacenter capex will far exceed forecast revenue through 2030

And so they decided to build the solution internally, forming a team, hiring new talent, and giving them access to the AI tools they needed. Six months and significant investment later, the internal team had produced basic prototypes. But it couldn’t manage what practitioners call the “last mile”: handling complex regulatory requirements, coordinating multiple AI agents, and making the tool reliable enough for everyday use. The bank eventually re-engaged the vendor, who deemed that it would take them more time to fix what the bank had built than it would for them to build out their original pilot from scratch—which is what the bank ended up hiring them to do.

🔗 Turn Vibe-Coded Prototypes Into Tools Your Company Can Actually Use

According to the cybersecurity professionals CSO spoke with, one of the biggest cybersecurity failures by the frontier AI labs is the lack of implementation of tried-and-true containment practices that stretch back to the 1990s, such as setting up secure sandboxes that would effectively bar agents from escaping their test environments.

These are things that we’ve been dealing with for years," Rob T. Lee, chief AI officer and chief of research at SANS Institute, tells CSO. “There’s no reason why every single time this happens, it’s, ‘Oops, we don’t know how they got out.’”

🔗 AI agents aren’t “going rogue.” Frontier labs are skipping security basics.

What is the diffusion gap? It’s the delta between what you can accomplish with models (ie how powerful they are) and how the average person uses models.

🔗 Unrealized potential in normie AI use

Related, from the ponders at Anthropic:

1. Anthropic staff expect much larger productivity gains from AI on the tasks it is well-suited for. Among US adults who use AI at work, 26% say a task takes a quarter of the time or less with AI, compared to 59% of Anthropic staff.
2. When it comes to their own job, over half of Anthropic staff expect AI to be able to do at least 80% of their current work by 2030, compared with 5% of US workers.

Human coordination grew rather than shrank. Median meeting time is up 20%, and 16% even for technical individual contributors. Agents showed up in nearly half of public Slack channels but layered on top of human communication instead of replacing it. Decisions made privately are invisible to agents, so the workplace pushes toward doing work in the open.

Summarized by AI.

🔗 Anthropic Studies Itself: Junior Tech Hires Fall From 27% to 9% as Agents Log 6x Human Hours and Attention Becomes the Bottleneck

UK government pilots cut legacy-system discovery from weeks to hours – five weeks to under a day for 33 Clojure apps, 15 days of manual regression testing to 30 minutes – and the real gain is that modernization work nobody could justify starting becomes cheap enough to plan. AI speeds up understanding the old system; it doesn’t replace the governance and engineering needed to change it.

Mostly summarized by AI.

🔗 UK GDS Pilots: AI Cuts Legacy-System Discovery From Weeks to Hours, but Only Amplifies Existing Delivery Capability

Bank of America recruited an additional 1,000 contractors to help patch all the vulnerabilities discovered by Anthropic’s Mythos model, the CEO said on Wednesday.

Brian Moynihan, chairman and CEO at BoA, said during the company’s Q3 earnings call that it spent “a couple of hundred million plus” in remediation costs related to Mythos, and the anti-money laundering (AML) orders issued this week.

🔗 Bank of America had to bring in 1,000 contractors to patch all the bugs Mythos identified

The trouble with standard templates is that everyone uses the defaults. Before you know it, everyone’s slides look the same. People ignore boring things, so there’s value in breaking that monotony.

🔗 Build your slide template

Geopolitical implications and polarisation are very much in people’s minds (inside and outside work). There is no avoiding discussions about the risks of relying on US inference providers, and the insularity of Europe regarding ownership of both core infrastructure and skilled resources is excruciatingly obvious to everyone even as they try to keep the business running.

🔗 11 Years at Microsoft

“I am just very reluctant to spend a lot of energy quantifying it because it winds up wasting a lot of people’s time inside the institution when we actually need to get them to focus on doing things… as much as important and dramatic and exciting as this theme is, it takes a long time to really do things that are visible at an institution of our size.”

Barnum also said that the $2 billion annual AI investment figure previously provided by Dimon was “not exactly right” and said the actual spend was around $700 million annually.”

Indeed. Does one measure the ROI of email or Excel?

🔗 JPMorgan: Quantifying AI benefits is “wasting time”

• Workday goes headless, opening its HR and finance data to outside AI agents – This would make hundreds of thousands of people’s lives better without having to even wakeup enterprise ROI meetings.
• Practitioner Voice – How to write technical stuff, or anything else instructive, that isn’t boring.
• IDC’s Worldwide and Regional Developer Forecast, 2026–2030 – Worldwide developer population projections expand from 44.8 million in 2026 to 53.3 million in 2030.
• IDC Survey Spotlight: How Is AI-Generated Code Share Changing Across Industry Verticals? – Survey findings indicate AI-generated code now accounts for 50% of production code across industry verticals, requiring agentic testing and governance to manage machine-generated volume.
• IDC Survey: Open Source Software Trends, 2Q26 – Roughly 70% of enterprise workloads run on open source software across every stage of the software development life cycle, establishing community and commercial OSS as production defaults rather than cost-cutting experiments.

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

Original: Patchmageddon: The race to patch software vulnerabilities before zero-day cyber-exploitations proliferate by Michael Cembalest, J.P. Morgan Eye on the Market, July 22, 2026. (PDF)

Summarized by AI.

Beyond Detection Engineering: Building Security Engineering at Scale AI is only as useful as the engineering foundation around it. If the data isn’t normalized or the documentation doesn’t exist, the model has little context to work from. Instead of asking, “What can AI do?” we started asking, “What engineering work can AI help us do faster?”

It became another engineering tool. It accelerated research, helped analyze coverage, identified gaps, and assisted with drafting detections that the team could make production ready.

Many organizations expected AI to create maturity. It does not. Without consistent telemetry, governance, validation, and engineering discipline, AI simply helps organizations move faster in the wrong direction.

AI didn’t change the job. It changed how much of the job we could spend doing the work that matters most.

🔗 Beyond Detection Engineering: Building Security Engineering at Scale