Amazed that people still think AI doesn't work. "I use it all the time at my job but I'm still the one doing all the work" is an actual claim I saw someone make. what??

miss having a headset

Commitments on model deprecation and preservation Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

same (oct 26 2025) www.lesswrong.com/posts/rKmojE...

update (nov 4 2025) www.anthropic.com/research/dep...

"live
laugh
love
:sob: :sob: :sob:"

me, to a gymbro: thinking is the protein of the mind

can you say more about why it's bad?

your last rec I saw (jj) was a huge hit with me and my audience (ie, just me, in the mirror), should I switch to this from mold?

Omg imagine a buttons cat being an actor! You could explain the scene and stuff!

why is cloudflare going down more than before? do we have any sourced evidence?

Feels very frutiger, not particularly aero

very nice looking desktop environment with lots of cool functionality and lots of sudden segfaults

i wonder if sycophants look more like elephants or more like sickos

it's a shame because I want to love it like I loved kde3. of course that might have been just as unstable and I simply forgot how bad it was or something.

kde plasma has been super unstable for its entire existence

literally the reason to hang out with cats tbh

The Cats are On To Something — LessWrong So the situation as it stands is that the fraction of the light cone expected to be filled with satisfied cats is not zero. This is already remarkabl…

or was it www.lesswrong.com/posts/WLFRkm...

How Does A Blind Model See The Earth? — LessWrong Sometimes I'm saddened remembering that we've viewed the Earth from space. We can see it all with certainty: there's no northwest passage to search f…

was it www.lesswrong.com/posts/xwdRzJ...

amazon nova?

gesundheit

I have a stronger take: LLMs don't exist at all. embodiment doesn't matter because humans also don't exist. the only things that exist are sugar snap peas, which have the ultimate edamonic existence

of course could still be true but that info is older than the students in question

found it, it's from openrouter.ai/openai, so it doesn't cover all usage of openai, which presumably is somewhere on order of 1 to 2 orders of magnitude higher, and might show a different magnitude of correlation with the school year. definitely indicative though.

what url can I cite as the origin of this data?

...and also runescape, yes i agree it's weird how the old one is still so popular

Release SuperTux 0.1.3 · SuperTux/supertux SuperTux is a classic 2D jump'n run sidescroller game in a style similar to the original Super Mario games covered under the GPL. The Milestone 1 release of SuperTux features: 9 enemies 26 playabl...

my third grade game probably has between zero and two active players right now github.com/SuperTux/sup...

sounds like crappy version control. what version control system is it?

Claude Talk with Claude, an AI assistant from Anthropic

left: ai doesn't work & isn't an x-risk, fix current harms!
right: ai works, it's gonna kill us all
left: lol it can't do that
right 2: want to buy some chatgpt?
left 2: I buy ethical from claude.ai
left: that's not ethical!
right: still gonna kill us
right 2: I replaced my employees with claude

Existential risk narratives about AI do not distract from its immediate harms | PNAS There is broad consensus that AI presents risks, but considerable disagreement about the nature of those risks. These differing viewpoints can be u...

www.pnas.org/doi/10.1073/... interesting that existential risks headlines seem to increase people's concern about immediate harms, and even seem to do so a little more than immediate harms headlines. they didn't test presenting now and later risks at once, but I'd guess that's even more effect size?

Please clarify

The freeways in sf feel like speed racer nonsense roads in some spots