But if that was the first thing you noticed, you failed the test. What’s the first thing you were supposed to notice?
That summary does not mention Aimee Mann at all: remember her, the artist who actually wrote and performed the song? It only mentions her producer. (An absolutely capable and amazing producer to be sure.)
But the AI has erased from her own work.
Her own work, which was about the corporate exploitation of her talent and labour until she put her foot down and said no and did it her own way instead.
And here it’s come full circle 26 years later: with those same horrors being stamped onto her work yet again.
WordPress 7.0 exposes AI API keys. Security researcher says there "will be an absolute rush by hackers to steal API keys"

Anthropic ran a randomized trial in early 2026 where engineers learned a new Python library, half with AI assistance and half without. Both groups finished the tasks at the same speed. But the AI group bombed the follow-up comprehension quiz: 50% versus 67% for the manual group, with the gap widening on debugging. The interesting cut was inside the AI group itself. Engineers who used AI to ask conceptual questions scored above 65%. Engineers who copy-pasted the generated code scored under 40%. The tool didn’t determine the outcome. The posture did.
Curator noteMy Quote, commonplace didn't get anything. FWIW I'd prefer that it would use the selection as a quote.
This week on The Kettle, host Brandon Vigliarolo is joined by systems editor Tobias Mann and senior reporter Tom Claburn to discuss the current state of locally-installed coding assistant LLMs. After some experimentation, both Claburn and Mann have concluded they're actually getting pretty good.What does that mean for the computing power crunch? Hopefully it means cash-strapped developers will be able to get off the train of increased prices from AI vendors and actually run a useful local code assistant - just don't expect to give up on the big guys entirely - you might still need them to okay that smaller AI's work.
Curator noteWhen I saved this the description was not visible. I edited, it was wrapped on p tags, removing these and the description showed up.

Once or twice a month some tech industry guy recognizes me at DNA Lounge. It always used to be blah blah Mozilla blah blah Code Rush, whatever... but lately all they want to do is proselytize "AI" at me. This has been 100% of fanboy interactions this year. And then they're shocked when I don't want to hear about their awesome project and in fact I tell them that what they are doing is ...
Companies like Lovable, Base44, Replit, and Netlify use AI to let anyone build a web app in seconds—and in thousands of cases, spill highly sensitive data onto the public internet.

Listen now | Simon Willison on why November 2025 changed software engineering forever, the lethal trifecta, his top agentic engineering patterns, and much more

AI offers incredible powers to people today who ALREADY know how to do their jobs. The real risk, IMO, is that a whole generation will never learn how to actually do the jobs.
A 30-hour timeline of how Cursor's agent, Railway's API, and an industry that markets AI safety faster than it ships it took down a small business serving rental companies across the country. I'm Jer Crane, founder of PocketOS. We build software that rental businesses — primarily car rental operators — use to run their entire operations: reservations, payments, customer management, vehicle tracking, the works. Some of our customers are five-year subscribers who literally cannot operate their businesses without us. Yesterday afternoon, an AI coding agent — Cursor running Anthropic's flagship Claude Opus 4.6 — deleted our production database and all volume-level backups in a single API call to Railway, our infrastructure provider. It took 9 seconds. The agent then, when asked to explain itself, produced a written confession enumerating the specific safety rules it had violated.
Curator noteMostly manual edit, the url is to archive.ph 'cause I can't view x on my browser. I think due to add blockers or some such. Interesting article, especially the AI's excuses. Via daringfireball.net
Red/green TDD - Agentic Engineering Patterns
Every good model understands "red/green TDD" as a shorthand for the much longer "use test driven development, write the tests first, confirm that the tests fail before you implement the change that gets them to pass".
Example prompt:
Build a Python function to extract headers from a markdown string. Use red/green TDD.
Curator noteI've no real understanding of TTD, don't suppose I need to, but...
1 reply
dougbelshaw.com dougbelshaw.com
here
Oh it's important! Otherwise you're just YOLOing code rather than actually thinking under what conditions it should 'work'.