RT AI Engineerbtw get to the Expo EARLY to get this sota swaglimited run while stocks lastbasically this is our equivalent of a magic 8 ball…
RT David HoangWho is coming to @aiDotEngineer World’s Fair in San Francisco next week?
RT Charlie O'NeillAn interesting question to think about is whether the labs will continual learn one big LLM or have personalised ones. If you did try…
Given same-ish margins, V4-Flash could be served at 175 t/s for ≈3x current speed, so $0,84/1Mt. Worth it? I doubtV4-Pro already hits 100 tps for me…
Idea: FujoshiBenchRussian has gendered verb conjugation. By default LLMs treat me as male. But here, a brainrotten Q (scarcity of fandom sexualization of Envy from FMA)…
RT MiniMax (official)next week at @aiDotEngineer, we are joining @togethercompute for a conversation on what goes into running agents at scale.@olive_jy_song, Research Lead, RL at MiniMax,…
RT Syd SteyerhartEveryone tells me that Elon Musk is a fascist because he could "solve world hunger" for a mere 3 billion dollars, but MacKenzie Bezos…
RT Tesla Owners Silicon Valley"To anyone I have offended, I just want to say I reinvented electric cars and I am sending people to Mars on…
RT Khurram JavedIt's interesting that people who are less in the weeds, like Dwarkesh, see the argument for continual learning better than people at the frontier…
Annoying that OpenAI doesn’t seem to give a GDPval measure for GPT 5.6. One of the best measures of economically valuable work.
impromptu ai engineer preshow floor tour and AMA https://x.com/i/broadcasts/1OxwbbVdAbDJB
DeepSeek being so keen on roleplaying will never not be funnyWhale V4.2 might become first love for many
RT Abhishek BhardwajI’m speaking at AIE SF World’s Fair on July 1st in the Sandbox & Platform Engineering track.My talk is called "From fork() to Fleet:…
weebs are mentally illMichael(Poltfan)((JapanDayTripper)): Now realize China does this with literally every aspect of their country. Literally every part.
✨infinite scroll✨ on Recent Chats is now available on our mobile app ^_−☆you can revisit every single chat you've ever had... even the ones from years…
it's a bit pitiful to look atEuropeans don't have very many things that China would pay immense money for. And these are specifically things their Master…
RT Leroy LiAnt以数倍于美国博士后时薪的价格向物理学家购买题目，因为数学家已经快被榨干了。但代价是这些题目不能作为研究对象，甚至无法拿来发论文，等于将知识产权卖给了Ant
An interesting way to take Noam at his word in regards to always keeping a constant inference budget for any eval reporting -is that open models…
RT xjdrwe've been running GLM 5.2 in bf16 and in fp8 (experts and kvcache only, attention is always bf16) and have recorded virtually 0 measurable quality…
RT Alexander DoriaI mean just HF out, ModelScope blocked and it’s over. You won’t seed GLM.Naithan Jones: “THeY cAnT bAn OpEn sOuRcE bRO hOw cAn tHeY…
RT Matthew Zirwas, MDThe take that frontier models aren't ready for use in medicine is dead wrong.I've been using Opus 4.x in my clinical workflow every…
RT yung macro 宏观年少传奇So just to summarize, the official US foreign policy posture toward Europe is:1. Overall, it is an intention of this admin to punish…
RT alphaXivSakana Fugu Technical ReportInstead of training one larger model, Sakana AI trains an orchestrator that reads each query and dynamically routes or composes GPT-5.5, Gemini-3.1-Pro,…
RT alphaXivSakana Fugu Technical ReportInstead of training one larger model, Sakana AI trains an orchestrator that reads each query and dynamically routes or composes GPT-5.5, Gemini-3.1-Pro,…