programming language idea: local variables are kiki and functions are bouba

we should have used fucking optocouplers, was the lesson learned last time, unfortunately the couplings can't be undone from here

is this distinct from being manic on our own? who can tell!

in that state where we're making the ground plane oscillate and it's giving the people we're coupled to a manic episode

you know what? fuck you *convolves your kernel*

#neverslop

still looking for a slop-free job! we have an incredible amount of skills in forward engineering and a reasonable amount of capacity for reversing, and we host our own email out of our house! US, remote, Pacific timezone. #neverslop #GetFediHired

donker inserting the "Gordon" LLM into their desktop app and im just

"AI will solve the simple problems so humans can focus on the hard ones" is such a ridiculous fail at understanding anything about how humans acquire skills and knowledge.

Humans HAVE TO solve simple problems and then advanced problems in order to learn how to solve the really hard ones. This is not optional.

Good luck solving any difficult problems in the future with a generation of people who spent their entire education typing every problem into a chatbot's prompt box.

Unclear, muddled thinking.
Redundant definitions.
Random lemmas that lead nowhere.
Unpublished results that nobody else understands.

It’s called “crackpot math”, and I was doing it long before LLMs, thank you very much.

@atax1a
I mean the newest rakes are so much better than the ones even three months ago.

us: stepping on rakes is a bad idea a priori

a bunch of fucking numpties with broken noses and metal tines in their feet: no you see you have to learn how to step on rakes productively, you can't criticize it unless youve been stepping on rakes yourself

I'd argue against calling software engineering "solved" (I'd use the word "disreputable" instead), but otherwise I agree with this article here.

It was already extremely disheartening to see basically the entire global software industry abandon all quality overnight. Imagine you run a restaurant out in the countryside and you have your regular customers, until one day some company puts up a dispenser on the other side of the road that pumps out some undefined, brown, lukewarm broth that tastes like cardboard but only costs 20% of a meal at your restaurant, and you have to watch as all but maybe one client never set foot in your restaurant ever again.

But it's much worse when it comes to things like blog posts. I have a list of technical topics that I once would've liked to do research on and write blog posts about... but what's the point now? I would only be feeding a bullshit generator so that it could generate some slightly more plausible bullshit in the future, making it yet more effort to debunk. And if any technical opportunity arose from it, then at least 3 slop projects would pop up immediately, in turn demotivating anyone who'd work on any reliable implementation (and their motivation had already taken a serious hit because they no longer consider the operating systems reliable that their software was supposed to run on, and it's a considerable amount of work just to remain where they are...).

And yeah, this year I've also noticed a distinct lack of... idk, "interesting technical things"? There was that one blog post from the FEX devs, that was really interesting. Then there was some good news about battery technology research (both lithium and sodium), which I assume will have a serious impact in the real world, but on an intellectual level, as someone who's not in that field at all, I'd file those under "neat". And that's it. That's the entirety of "interesting creative technological output" that I remember coming across in 2026.

https://borretti.me/article/no-man-is-an-island

just in case you didn’t know that “it’s only for code reviews, I don’t see the problem” is always bad faith horseshit: https://www.phoronix.com/news/Linus-Torvalds-Claude-Code

> Linux creator Linus Torvalds has been known to use AI to help with kernel debugging and outside of the kernel tree has also done some vibe coding with Google's Antigravity. He has previously said he views AI as a useful tool. It turns out he's even now using Claude Code for working on new Linux kernel features.

> "The code, the test programs, the review behind the list above and the first draft of this text were done with an LLM coding assistant (Claude Code) under my direction. The measurements were run on my machine."

so just a full vibe, done with the LLM that can’t make a web browser, C compiler, rsync features that aren’t riddled with bugs, or JavaScript runtime

this is the shit we were warning you about

was there a reason why you didn’t listen?

I still have no genuine use case for generative "AI". Not one.

Yes, I know what other people are using it for.

But none of those are things I would want to use it for, given its errors, (real) cost, side effects, ethics, and politics.

And this isn't an "oh, if they fix this one issue, I'd be onboard" kind of thing. They would have to fix ALL of them and they are all inherent to the product. None of them are even fixable.

sighing heavily as people go into the exeptionalist/nationalist trope of "russia bad" instead of reflecting on how american culture enables and is amenable to influence ops like that

wishing all american liberals to realize that the blame for our problems does not lie in the hands of russia. russia is certainly exploiting our problems, but it did not cause them. america did that on its own. america's problems are not caused by mysterious foreigners doing evil. america's problems are caused by americans choosing to align with evil.

so on OpenAI's math proofs, which are generating a pile of Lean that compiles and may or may not match the problem statement

mathematicians are already saying this isn't useful, even if it does match the problem statement, which is years of work to determine that fewer and fewer people are willing to do for blatant spam

i am not a mathematician, but I'm getting the idea these "proofs" aren't useful in any applicable sense at all. are there any that are?

had some dream stuff going on but we forget what exactly happened

Silicon Valley is falling in line with Trump’s attempt to rename AI - https://www.theverge.com/ai-artificial-intelligence/1009406/silicon-valley-is-falling-in-line-with-trumps-attempt-to-rename-ai it's a loyalty test: who is sycophantic enough to adopt a stupid name for #ai