@inthehands my, perhaps hot, take is that a tremendous amount of this conflict between clay-like and rigid computing is essentially an immune disorder in the response to fascism’s downstream effects.
Much of the difficulty of modifying a modern computer is because of the difficulty in distinguishing the user from the NSA, or from the organized crime that thrives in high-corruption low-quality-of-life environments.
@skinnylatte ever been to Angel’s Tijuana Tacos in LA? After visiting there I’ve been forlornly hoping I can find something on their level in the Bay Area, I should check out Tacostao
Gotta say I don't buy this argument. The red herring is the idea that the developers of the code are the only ones with an interest in it. Different code produces different results for end users, and we have a tremendous responsibility to them.
Focusing on readability is often (usually? always?) still the right thing to do, but it's *for* them.
Something I've noticed about performance work is that people really want there to be A Cause for improvements. A new technology, something old no longer holding things back, "removing the debug logging", etc…
The truth of the matter is that 97% of the time the enabling thing for performance work is engineers having time in the schedule to sit down with the tools (usually a sampling profiler) and grind out small wins over and over.
There's no magic bullet that replaces good engineering.
My code is read in earnest by maybe, let's say 10k people, and run by about 2 billion. So a shallow analysis would suggest that for this to be true, we need to derive about 200,000x more value from reading than from running.
However!
Many of those people who read it then go on to change parts of it, which creates a feedback effect where the 10k influence the 2B.
@inthehands exactly, yeah. That then turns into the debate of “is misrepresenting your level of certainty dishonest?”, which is imo a separate almost-unrelated issue from the original
My perhaps-unkind pet hypothesis on this is that a lot of people evaluate the world *primarily* through the lens of social judgement: the actual information content is irrelevant, the question is automatically and silently mentally restated as "did Maria do Wrong®?"/"should Maria be Judged® for saying this?"
LLMs generate statements that are not *known* to be true, which is different from them not being true. 2 + 2 is still 4 even if an LLM told you that.
This would mostly just be whining about grammar except that mixing up knowledge and truth actually has consequences for effectively conveying meaning in this case: people will interpret “not true” as meaning “false”, and then dismiss the argument as bad faith or ignorant after testing it.
Hi folks; this is something I've been hesitant to post about since it's a very sensitive topic for me, but:
I'm moving out of the US later this year, and unfortunately for legal reasons I'll be unable to take my dog with me. All the rehoming services I've contacted are unable to take her in, so, I've made a little page about adopting her.