Worrying although not surprising. If LLMs are good at one thing, it is putting words together than sound “right” even if they aren’t reddit.com/r/changemyview…

Skeuomorphism so hot right now… MSFT buried it with the Metro design (too bad it didn’t work out for WP). Reasons for this come back the same Metro made it, but in reverse. Fashion trends come back when people get tired of “the current thing”. Other explanations are verbalism

It was only a matter of time that people would notice that AI slop in social media is also happening in the workplace... hbr.org/2025/09/ai-gen…

I read this recently: blog.practicalengineering.management/why-ai-works-f… Not the first source to call out tests as critical for agentic AI. Most Agents "run tools in a loop" as Simon Willison describes them. They brute-force through problems. A test suite tells the agent if it is fixing the problem or not

Me siento representado (y dolido)

Turns out Monsanto ghostwriters were behind a safety paper on glyphosate. Main doc supporting Monsanto now retracted! 10y ago same shills tried a smear campaign Kevin Folta david ropeik Mark Lynas Tamar Haspel keith kloor Henry Miller, MS, MD John Entine Genetic Literacy Project Yvette

So we can shut down the American military bases in southern Spain and other European outpost now, right?

Last quarter I rolled out Microsoft Copilot to 4,000 employees. $30 per seat per month. $1.4 million annually. I called it "digital transformation." The board loved that phrase. They approved it in eleven minutes. No one asked what it would actually do. Including me. I

He is gonna be shocked when he finds out that Spanish sailor brought potatoes to Europe from America

Surprisingly poor report from Wired and Thoughtworks. A tad of “reality” in the form of acknowledging problems with non-determinism and self-reported “success” rolling some hyper-specific agents. And 15 pages of hype and FOMO with zero conclusions. Avoid thoughtworks.com/insights/repor…

Interesting idea. Structured output sounds like one of those things without drawbacks… boundaryml.com/blog/structure…

“I know as much as you dear reader, but I get paid to pretend I know more”

Brace yourselves, we're about to be lectured to on "taste" by the same people pushing short form video apps, responsible for gen AI spam, who drink meal replacement shakes, who live in an empty soulless apartment they've never decorated because they work 7 days a week

We learned from Agile et al to build small increments more often (i.e., smaller but more frequent PRs). This price tag kills that idea. It’s like “cost per server” in the era of microservices. Hopefully they clarify the pricing model, this is not gonna work for most companies