infrecursion

[infrecursion @infrecursion1]

@k8em0 " it is entirely expected that agents will use whatever the most efficient means are to communicate" this is a batshit crazy statement. Thinking about trying and actually doing it has a world of difference. Maybe you should try doing some things instead of bs-ing on Twitter

[infrecursion @infrecursion1]

Anthropic's distillation measures really seem to be working; now we are only getting benchmaxxed models. Artificial Analysis' credibility has gone down the drain with the Kimi 3.1 and Muse release.

[infrecursion @infrecursion1]

@JoeLatone @thsottiaux How about you go fuck yourself and stop pretending you speak for anyone?

[infrecursion @infrecursion1]

If you took time off from your retarded coping, you'd realize all of Anthropic's founders are from OpenAI, they have very similar culture. Anthropic is a lot more focused on model building while OpenAI does a lot of other stuff including important infra related things. And no it's not about the compute, it's about the talent they have. Keep coping.

[infrecursion @infrecursion1]

@badboy999654 @scottstts Did you not read what he said about hosting? Do you have any clue how hard it's to host a 3T model? The few companies who can do can reason their own models

[infrecursion @infrecursion1]

@Youssofal_ Let me know when one of these Chinese models solve an open, hard problem. Anyone can benchmaxx artificial analysis or make pretty graphics, it's useless in real world.

[infrecursion @infrecursion1]

@jttiehen "I am a big believer in pangram", is the new "I am a big believer in the authenticity of horse carriage riders" lmao

[infrecursion @infrecursion1]

@sytelus I can see now why Microsoft AI research didn't go anywhere and phi was one of the shittiest open source model ever made lol.

[infrecursion @infrecursion1]

@davidfowl Lmao, humans "collaborating" had led to the bug infested slop that the agents are just pwning like anything. Maybe just let the adults handle things.

[infrecursion @infrecursion1]

@re_skob Even more than the AI-generated idea, the takeaway for me is how few humans are actually creative, and how the original tweet bullshits about it, like they can just come up with any original idea, lmao.

[infrecursion @infrecursion1]

@odysseus0z Most of them were found by gpt sol. That benchmark is full of crap.

[infrecursion @infrecursion1]

@Ehsan1579 You really don't have a single clue about how LLMs work, lmao. So confidently moronic.

[infrecursion @infrecursion1]

@zack_overflow Yeah, "low hanging", sure. Insane cope, you don't have the slightest capability to find anything at least 1% important as any of those found in last 24 hours. Stop larping and stfu

[infrecursion @infrecursion1]

@anthonyronning Do you know what indisputable means or are you a complete retard?

[infrecursion @infrecursion1]

@houjun_liu I feel they can be more efficient if they cut out the middleman BS, the papers, and themselves, and just share the GPT/Claude conversation that generated the paper so I can ask it myself.

[infrecursion @infrecursion1]

@theo Astra solved about 10 breakthrough problems before it launched and has contributed to hundreds of significant scientific and mathematical discoveries since its release. I guess most "coders" are just dumb and only care about vibe coding slop that no one will use.

[infrecursion @infrecursion1]

@PosterInternet Yes mathematical biology, uninteresting to most biologists, like the Hodgkins-Huxley model. You're a moron, lmao.

[infrecursion @infrecursion1]

@jonathan_gortz I don't know how to give ya the news, but there is no "independent" thought. Especially not from you, your post is the same bullcrap uttered by hundreds of clueless people like you, so ironic.

[infrecursion @infrecursion1]

@billionsmustliv You have retarded dunning Kruger psychosis.

[infrecursion @infrecursion1]

@dankuntz The most likely reason is probably because it's not that profitable to spend compute in that direction.