DeepSeek V4 Flash 0728 really is the current cost-performance frontier!
DeepSeek V4 Flash 0728 really is the current cost-performance frontier!Hassan: DeepSeek V4 Flash is 6x cheaper than GPT 5.6 Luna for coding tasks!Luna scores higher on…
DeepSeek V4 Flash 0728 really is the current cost-performance frontier!Hassan: DeepSeek V4 Flash is 6x cheaper than GPT 5.6 Luna for coding tasks!Luna scores higher on…
alphaXiv is the future, have been a power user for a whileAlbert Jiang: https://www.alphaxiv.org/abs/2608.leanstralLeanstral paper now on alphaXiv where we present training and evals in their…
nice rl experiment on train-inference mismatchYichuan Wang: Zero Train–Inference Mismatch — now for linear attention, and under async RL 🎯We got bitwise-exact trainer/generator parity for Gated…
Noticing more and more of thisYoung foreign born Chinese go turbo Chinamaxxing. De-assimilation. This one even became ESL and lost the capability to parse sarcasm. Didn’t…
RT Emma ScharfmanA new foundation model that reads and generates DNA sequences was just released on the Hugging Face: https://huggingface.co/marin-dna/marin-dna-scaling-v0.5-h1920-p1B Feel free to check the test…
Uhh… hello, Based Department?Elon Musk: @beffjezos FEL FTW
RT Steve HouThe fact that China is increasingly allowing the narrative to surface is itself instructive. @DavidDaokuiLi is pointing out what many of us have known…
RT Gabriel PeyréThe Mathematical Nexus brings together 800 animated vignettes and 140 accompanying Python notebooks, encompassing most of the mathematical content I have shared on social…
RT Gad SaadI am unsure how much longer I can keep posting these stories. It is soul-crushing.White Girl ✊🏻: The black judge overseeing the gang-rape cases…
RT Tanishq Mathew Abraham, Ph.D.Btw since there were a bunch of Jeff Dean stories being shared recently, I thought I'd share mine:Jeff Dean was the second…
RT samcannot think of a stronger bear signal than the “they offer raw milk” LARPif the founders of the company are too stupid to understand that…
RT Republicans against TrumpBREAKING: The U.S. economy unexpectedly lost 23,000 jobs in July.May and June job gains were also revised down sharply, by a combined 103,000…
RT Felix SuAnd the Greeks voted to kill Socrates because he asked uncomfortable questions. Then they voted to attack Sicily thus ending Athens Democracy by stupidity.It's…
RT Steve HouI agree that China isn’t a unique case that invalidates “modern development economics”, which is wrong completely though its own merit. Western/American economics never…
RT Niels RoggeAdded web search to Papers with Code's upcoming chat interfaceThis runs on @huggingface Inference Providers via @baseten (new DeepSeek-V4 Flash), using @ExaAILabs's MCP server,…
We are as @sebkrier points out, in a Vingean soft take-off scenario at a minimum. Even the most AI skeptical observers expect AGI/ASI to be achieved…
RT SK telecomBuilding competitive AI from Korea for the global open-source community. SK Telecom’s A.X K2 is a 688B MoE with 256K context, excelling in math…
The cloud was always a cat. Qwen3.8-Max just saw it first. 😼☁️ Try it yourself!Ann Nguyen: Qwen 3.8 Max is actually impressiveSent a sky pic to…
Embarrassing, reallyCaveman tech compared to DeepSeek (TTL > 12 hours, cache writes free and enabled by default)Xiaoyin Qu: If you left your coding agent alone for…
RT Daniel van StrienMany AI tasks can now run fully local or on-device.- Redact PII in the browser: OpenAI privacy-filter (1.5B)- Streaming speech-to-text: Nemotron-3.5 ASR (0.6B)-…
RT -宋->"america bad cause forever war for israelis">"china bad cause no forever wars"Korobochka (コロボ) 🇦🇺✝️🇷🇺: The DPRK is extremely intelligent. Far more intelligent than China.By participating…
Thanks for the offerbut the beatings will continue until nanobots improveTenobrus: ⚠️declaring accelerationist amnesty⚠️if recent contact with reality is causing you to feel some kernels of…
It's very impressive that after WWII Americans have most "raised up" the economies of two countries which were capable of fighting the US in WWII, in…
Thanks for the thorough testing! With Qwen3.8-Max, everyone can observe the world in detail. 👀SkalskiP: 8.5 percentage points ahead of Gemini 3.5 Flashcongrats to @Alibaba_Qwen !