Category: tech
-
What Broadcom said about using Mythos at Explore 2026
Broadcom has been using Mythos to find and address security vulnerablities since the begining. Here are recent comments about that from Explore 2026.
-
AI is valuable to the individual, less so to the organization
Ask an individual if AI made them more productive and about four in five say yes. Ask an organization if it moved EBIT and it’s fewer than two in five, and one in sixteen for anything you’d call real. A chart pulling three August surveys together, plus why the goal should be getting better at…
-
One CEO in eight got what AI was sold as, survey
PwC asked 4,454 CEOs what AI did to their revenue and costs over the last twelve months. The biggest single box on the matrix, at 42%, is the one where nothing moved. One in eight got both benefits at once, and 13% came out worse.
-
Using AI to play Role Playing Games – Building tools to play Dungeons and Dragons is playing D&D…?
I wanted to see how much time I spend playing D&D with AI versus building the tooling for playing D&D with AI. Charted across 4,245 conversations back to 2023: harness building goes from 5% to 53%. The cobbler’s curse – you love making shoes so much you never walk around in them.
-
Platform as a Product, 9 years later
Nine years after the platform as a product paper first came out, Caleb Washburn and I went through what changed. The answer is mostly “not much,” and the good bit is why developers keep asking for Kubernetes clusters they don’t actually want.
-
Using frontier models to find security problems in Spring Framework – a trickle turned into a tidal wave
“In March, we saw 55 reports coming from the community. In April, that went to 112. April is also when we finally got access to those frontier models and started doing internal scanning, and our internal scans gave us another 370 internal reports.” Michael Minella on the Spring Team. The historical baseline was about seven…
-
How I tell AI to write in my style, other people’s styles – and how to strip out shitty AI talk and LinkedIn language
An agent skill to remove AI writing tells, LinkedIn style, and write like Coté.
-
How people use AI in 2026: TV, homework, the DMV
Google looked at 15 million Gemini conversations and found people mostly asking about TV, homework, and government paperwork. Work is 13.5% of it, and almost none of that is trying to automate anything. Also: what the study leaves out, which is most of enterprise AI.
-
AI won’t fix your broken culture, but you can fix your broken culture
AI amplifies what you have: good processes get faster, bad ones get worse. Getting real enterprise AI ROI means changing how you work – structure, focus, scope.
-
Normie AI Usage: Figures & Findings
Programmers are still the dominant users of UI and, thus, drive AI use cases. The much larger market are “normies,” people outside of the IT department who are doing “knowledge work” and other tasks in organizations. I asked the AI to pull together a list of what those activities are and the momentum around them.…
-
What people use sovereign cloud for, and why
IDC survey data on sovereign cloud: more than half of organizations already use it, 72% say their interest has risen in the past six months, and 42% are now running AI workloads on sovereign infrastructure. Here’s what they’re using it for, what’s driving adoption, and where current offerings still fall short.
-
Platform as a Product, 2017 to 2026 – not much changed, which is the point
The Platform as a Product paper, first published in 2017 and updated for 2026. The concept survived Pivotal, the platform engineering category, and kubernetes blocking out the sun for five years.
-
How AI change your software development organization
A catch-up with Purnima Padmanabhan, GM of the Tanzu Division at Broadcom, on what her team has actually learned shipping enterprise software with AI for the last year and a half: code generation is the small part, beautiful code is the new uncanny valley, and you cannot solve the agent boundary problem from inside the…
-
Enterprise self-harm: cleaning the data is the hard part
I think the critical part of it was really realizing that we had built the original product presupposing that our customers had data integrated, that we could focus on the analytics that came subsequent to having your data integrated. I feel like that founding trauma was realizing that actually everyone claims that their data is…
-
50%+ failure is normal
Analyst firm Gartner thinks at least half of all generative AI projects “will overrun their budgeted costs due to poor architectural choices and lack of operational know-how,” and most organizations that try to build custom models “will abandon their efforts due to costs, complexity and technical debt in their deployments.” Yes, and this matches decades…
-
Always taste the digital transformation while you’re making it
At a large enterprise I recently worked with, the board asked the Chief Digital Transformation Officer to develop an AI adoption strategy to drive innovation, growth, and cost efficiency. His consultant of choice conducted stakeholder interviews and proposed a three-phase program scheduled to last 3.5 years: Phase 1: Fix digital basics and address the leftover…
-
Software Defined Talk #122: Emily Long, VCs, startups, Edera, etc.
In this episode, Whitney and I talked with Emily Long, CEO and co-founder of Edera about their hardened container runtime – the kind you can swap in without re-platforming or kicking off yet another zero-trust migration. The pitch is basically: stop chasing detect-and-respond alerts and just fix the foundation. We also got into the messy…
-
Let them tinker – hacking developer resistance to sound enterprise architecture and platforms
Developers need to tinker or they’ll reject your platform. That is a lesson that people who build tools and platforms for developers learn. The more ambitious your platform is in scale, the more tinker resistance you encounter – you want it to be the platform that 10,000’s of developers at a bank use, for example.…
-
Using a custom CLAUDE_CODE_DIR does not work
I’ve tried to use a different $CLAUDE_CONFIG_DIR for several weeks now. This way, I can have partitioned off harnesses for real life stuff, versus code, versus experiments. It doesn’t work, however. Sometimes Claude obeys it, sometimes it doesn’t. This means the wrong permissions, memories, rules, and ways of operating get loaded and mixed in. You…
-
If you can’t wait until Friday for the perfectly edited and polished Software Defined Talk podcast episode, you can watch the unedited, full recoding – in glorious color video – in the meantime.
You must be logged in to post a comment.