"hypotheticals are not an unalloyed good" i shout while walking into the Temple of Pure Reason.

This weird thing where people guffaw at the "conspiracy theorists" for caring that basic institutions are corrupt and coopted, like it's cringe to think not everything is perfect and "democracy" is fine on autopilot.

LLMs are civilization-tier muscle memory.

Btw the reason why computer-wide stutters went away in the multi-core age is because programmers are exceedingly bad at saturating multiple cores at once.

In my experience this is usually a stealth comment on the American style of commercial restaurant dinner service, where they try to chase you away as soon as you're done eating, so as to open up a new table. Same with starting to eat before everyone has their food. This is what.

Endlessly annoyed that image searching for vintage pictures now gets you vintage-looking AI slop.

"Model-View-Catharsis". For those who are not ready to embrace the future yet, a tale of why the past must be abandoned, and was never real to begin with. Features bold and experimental writing. ! Actually it's my most horny-on-main post tbh.

"The Incremental Machine". A glorious dream about an incremental coding future that cannot possibly ever happen, except that I built Use.GPU exactly how I said I would, and it works today.

"I is for Intent". Declarative UI won and is now a Pure Function™ of State. But is it really?. Actually it's the opposite. Most people don't quite know how it's supposed to be done at all, and build byzantine horrors in a React-like they hate and mistrust.

"On Variance and Extensibility". Learning about types is one thing. Learning how to apply the theory to applications is quite another. There are decades of past computer engineering mistakes all around us, but have you ever looked closely?.

"APIs are About Policy". Forget everything you think you know about web APIs. They lied to you. Instead of arguing about REST vs GraphQL, understand why both are really quite terrible, and in your way.

Thread of evergreen acko posts for the discerning gentleman and fair lady of the hackercraft persuasion.

RT @Fremond_: I’m in a really bad place rn. Not mentally, just the European Parliament

This so perfectly captures the specific kind of cringe associated with scenarios in academic and AI research. It's like the people involved don't understand writing at all. Zero. Nada. They are unable to step outside themselves and do the most basic sanity check about what they

RT @EricRichards22: The zoomers don't even believe the world that was taken from them existrd.

Maybe it's time for a classic "Falsehoods people believe about AI assistants". - That the AI will give the same, consistent response to the same question multiple times, so you only need to ask once. - That the AI learns over time from interacting with you, in particular after.

RT @katrosenfield: here it is, my contribution to the discourse

Writing a best-selling book about a political group: propagandist. 😡😡😡. Writing academic papers about people you have a grudge against: scholarship! 🧠💡👈.

This talk had some very impressive codebase-wide dev tooling on display.

ok but what if