Tri Dao @tri_dao
Asst. Prof @PrincetonCS, Chief Scientist @togethercompute. Machine learning & systems. tridao.me Stanford, CA Joined May 2012-
Tweets1K
-
Followers46K
-
Following667
-
Likes2K
.@nvidia Nemotron 3.5 Super VL is now live on Together AI. An open-weight multimodal model built for visual reasoning and always-on agents that see, reason, and act across documents, screens, and video.
TPUs have different bottlenecks from GPUs: few cores so you don’t need yo think about parallelism as much, large VMEM so easier to fuse kernels, but vector units have lower throughput so anything non matmul require much more care
Releasing FLAT: Flash Linear Attention on TPUs 🚀 Pallas training kernels for Linear Attention, Mamba-2 and GLA. On one TPU v6e, forward + backward is up to 3x faster than FLA on an H100 for GLA, and 1.5x faster for Mamba-2. 🧵
[1/4] 🧵 Mamba-3 rotates its recurrent state. What does that buy beyond retention? Our new paper connects these rotations to underdamped second-order dynamics: decay controls the damping envelope, while phase shapes oscillation. 📄 arxiv.org/abs/2610.13341
Great work from @zwljohnny, @tri_dao, and @nvidia on ReplaySSM! Mamba2 standard decode and FlashInfer-backed MTP support are already in vLLM, reducing recurrent-state memory traffic by caching SSM inputs and updating the checkpoint only when needed.
ReplaySSM now runs in production in GLM-5.3-Flash and Kimi-K3, and ships in TensorRT-LLM, vLLM, FlashInfer, and SGLang! It speeds up standard decoding (up to 1.47x) and speculative decoding (up to 2.1x), and now also enables a low-precision SSM state whose rounding error never
transformer inference has had dozens of tricks optimized by the entire ecosystem, there are still many ideas and optimizations waiting to be discovered for alternative architectures!
ReplaySSM now runs in production in GLM-5.3-Flash and Kimi-K3, and ships in TensorRT-LLM, vLLM, FlashInfer, and SGLang! It speeds up standard decoding (up to 1.47x) and speculative decoding (up to 2.1x), and now also enables a low-precision SSM state whose rounding error never
SSMs had the nice asymptotics but kept missing out on the inference goodies Transformers get for free like spec decode and quantization. Turns out we were just not caching the right thing. Same IO-aware trick behind FlashAttention, Mamba scan, paged KV. A lot of the big wins weren’t new architectures, just people taking memory IO seriously.
ReplaySSM now runs in production in GLM-5.3-Flash and Kimi-K3, and ships in TensorRT-LLM, vLLM, FlashInfer, and SGLang! It speeds up standard decoding (up to 1.47x) and speculative decoding (up to 2.1x), and now also enables a low-precision SSM state whose rounding error never
Together is the #1 provider by volume of tokens (across closed and open endpoints) on @OpenRouter openrouter.ai/providers
Announcing Together Link! Run frontier open models in your favorite coding harness. Track your spending and let our Auto router pick the model for you: low-cost models for quick fixes, more capable models for harder tasks.
We're releasing tev1-4B-experimental, a Jev-like classifier finetuned on top of Qwen3.5 4B. We're making it available on Together serverless at $0.042/M input & $0/M output. Also releasing the data recipe & a tutorial on how to finetune your own (Tev1 cost $17 to train!).
we got early access and hipkittens is now running on amd helios mi455! here we share an educational gemm ladder to demonstrate how to use the new helios features. overall, we found that the patterns identified in hipkittens for writing performant mi350/355 kernels translate well to helios making the forward port quite smooth!
Mayank put in a crazy amount of work to get pretraining to work on 3 gens of Nvidia GPUs and 2 gens of TPUs! Very good model for such a small size
We pretrained a 2.3B MoE (360M active) Hybrid Mamba-2 that lands within a few points of Llama-3.2-3B using <1% of its pretraining FLOPs. No dedicated cluster. The run hopped between H100s, A100s, V100s (yes, V100s) and TPU v5p/v6e on a single codebase. Meet Rigel 🧵
A lot more gpus coming for open models
We just signed one of the largest AI infrastructure deals for open source, period. 250MW data center. $5B+ in annualized revenue. Built with @HUMAIN in Saudi Arabia. tinyurl.com/4nzcjcu7
These guys move fast, 1st rack already ships. More inference compute is always welcome
We've raised $700M at a $21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z, Peter Thiel, BCV, and Blackstone. We're also excited to share that we've shipped our first rack to Jane Street.
I really like this new benchmark. Has the flavor of ARC-AGI3 but it’s pure text so you don’t have to worry about the vision capability
We’re excited to announce DiG-bench, a new benchmark for discovery! Over the last few weeks we’ve been testing frontier AI models on our novel discovery games and seeing how they score. Each game is a text-based environment, so they probe discovery capabilities in the natural
Only took 6 years 👀 github.com/NVlabs/CuTe
the gains come entirely from wrapping π0.5 in an agent that plans, verifies, and recovers. no new finetuning. @lianegalanti and team make the case that "agents for robots" can narrow capability gaps in off-the-shelf robot policies.
Robot policies can move but can't think. LLMs can think but can't move. So we connected them. Real robot: 16.7% → 97.3% Sim (LIBERO-PRO): 12.8% → 53.3%
Congrats to this super ambitious team
Today we’re announcing our Series B. We’ve raised $200M at a $2B valuation from Greenoaks with participation from Index Ventures, Hanabi, A*, Bain Capital Ventures, CVS Health Ventures, and Definition. Our mission is to simulate all eight billion people on earth, accurately.
Putting LLM brain on robots -> 4x SOTA with no extra training. I’ve been very surprised by how well this works. The time for agents running on robots is coming soon
Robot policies can move but can't think. LLMs can think but can't move. So we connected them. Real robot: 16.7% → 97.3% Sim (LIBERO-PRO): 12.8% → 53.3%
Kimi K3 is now live on Together AI. We’re proud to be a Day 0 launch partner for @Kimi_Moonshot’s open frontier model, built for long-running agentic workflows across code, tools, vision, and research.
Congrats to this team, really strong on both hardware design and kernels
We’ve raised $300M in Series C funding at a $10.3B valuation from Sequoia, Andreessen Horowitz, Jane Street, Argo, and SK Hynix. Our mission is to run the world's inference. This round accelerates production of our inference clusters. We've opened an 80,000-sqft, 10-MW facility
Andrej Karpathy @karpathy
4.4M Followers 1K Following
Jim Fan @DrJimFan
600K Followers 3K Following NVIDIA Director of Robotics & Distinguished Scientist. Co-Lead of GEAR lab. Solving Physical AGI, one motor at a time. Stanford Ph.D. OpenAI's 1st intern.
Percy Liang @percyliang
127K Followers 424 Following professor of computer science @Stanford @stanfordnlp, co-founder of @togethercompute, https://t.co/7R5THVnJ6u, @simile_ai, pianist
Aran Komatsuzaki @arankomatsuzaki
196K Followers 442 Following Sharing AI research. Early work on AI (GPT-J, scaling, MoE). Ex ML PhD (GT) & Google.
Horace He @cHHillee
55K Followers 620 Following @thinkymachines Formerly @PyTorch "My learning style is Horace twitter threads" - @typedfemale
Lucas Beyer (bl16) @giffmana
155K Followers 651 Following Researcher (now: Meta. ex: OpenAI, DeepMind, Brain, RWTH Aachen), Gamer, Hacker, Belgian. Anon feedback: https://t.co/xe2XUqkKit ✗DMs → email
Mark Tenenholtz @marktenenholtz
155K Followers 729 Following Member of Technical Staff @perplexity_ai
Tim Dettmers @Tim_Dettmers
49K Followers 918 Following Creator of bitsandbytes. Professor @CarnegieMellon and Research Scientist @allen_ai . I blog about deep learning and PhD life at https://t.co/Y78KDJJFE7.
Soumith Chintala @soumithchintala
329K Followers 1K Following Building new things @thinkymachines. Also dabble in robotics at NYU. Cofounded @PyTorch. AI is delicious when it is accessible and open-source.
elvis @omarsar0
325K Followers 1K Following Building the frontier @dair_ai • Prev: Meta AI | PaperswithCode | Elastic | PhD
Eric Jang @ericjang11
145K Followers 4K Following
Gabriel Peyré @gabrielpeyre
102K Followers 444 Following @CNRS researcher at @ENS_ULM. One tweet a day on computational mathematics.
typedfemale @typedfemale
49K Followers 566 Following a really exciting new account "advanced pytorch user" - @cHHillee alt: @typedalt
Delip Rao @deliprao
73K Followers 5K Following AI researcher @Penn. Past: @johnshopkins, @UCSC, @Amazon, @Twitter. alt: 道元, improv, running 🌈
Beidi Chen @BeidiChen
17K Followers 412 Following Asst. Prof @CarnegieMellon, @amazon Scholar, Prev: Visiting Researcher @Meta, Postdoc @Stanford, Ph.D. @RiceUniversity, Large-Scale ML, a fan of Dota2.
Richard Socher @RichardSocher
124K Followers 1K Following Building self-improving superintelligence CEO @recursive_si and @youdotcom MP @aixventuresHQ Ex: Stanford Adj Prof, Chief Scientist at Salesforce, CEO MetaMind
Ross Wightman @wightmanr
25K Followers 1K Following AI / Computer Vision. Ex 🤗. Ex head of Software, Firmware Engineering at a Canadian 🦄. Building AI systems or investing in startups that do it better.
Alessandro `Scinawa' ... @scinawa
1K Followers 4K Following Researcher in quantum algorithms @quantumlah. Entrepreneur. European. β \to +∞.
Shawn Wei Chew @sw_5405
1 Followers 49 Following
kpjvwdqx @kpjvwdqx
0 Followers 41 Following
Anar Rzayev @AnarSnowball
472 Followers 2K Following Cooking proteins with ML @ ISTA 🇦🇹 Ex-Research student @ KAIST 🇰🇷 | Prev CS & MATH @ EPFL 🇨🇭 | Bronze medals @ IMO 2017, 2019 🇧🇷 🇬🇧
Sdlm @Sdlm334287
41 Followers 1K Following
Caped0j @Caped0j
31 Followers 2K Following
Chetan Kulhari @ckxyz_
1K Followers 489 Following cofounder @midcenturyai | prev @stanford, @ycombinator
Boris @BorisMeinardus
1K Followers 161 Following 👨🏻💻Research Scientist @RadicalNumerics, Ex-SakanaAI 🤖 I like building AI and sharing my knowledge and experiences!
sioelnkc @v93ncofw0j9h
0 Followers 15 Following
Varun Ojha @vojha_nyu
2 Followers 36 Following Principal Software Engineer | Cloud & AI | Technology, Economics, Education & Global Markets
Otávio de Oliveira C... @PelegoCorrea
6 Followers 570 Following
ruinae @ruinaeruinae
23 Followers 619 Following Student @ CS Institute || Defence Tech & AI Enthusiast
Eduardo Garcia @eduasgarcia
29 Followers 120 Following trying to advance the inevitable ai takeover one hyperparameter at a time https://t.co/nHqUcgFmoD
Md Nadeem @mdnadeemai
122 Followers 529 Following Founder, CoEdify | Production software and AI for fintech, payments, insurance, and B2B SaaS | Built devsko + revsko, live products | ex-Oracle
rvrhms5 @b92xt
57 Followers 1K Following
manish @____manish___k
3 Followers 250 Following machine learning, optimization 365 days of machine learning. day1=16th October.
Robin Wang @WangRobin58334
43 Followers 327 Following PhD student @princeton in Robotics, prev @mondorobotics @umich
Wilbur Wu @wilburwuX
8 Followers 43 Following 20+ years in GPU/SoC performance architecture. Exploring AI agents, inference systems, and how AI can reshape hardware design.
Chia-chi Hsieh @CHIACHI_HSIEH
0 Followers 63 Following Master @Tsinghua_Uni | Intern @thukeg & @Zai_org | LLM serving & efficient AI
Rui Sun @Rs33me
8 Followers 118 Following AI researcher on agentic post-training & RL env. Prev: @Amazon / PhD @MIT / BS @Tsinghua
Akash Kannan @_AkashKannan_
2 Followers 108 Following
Ryo Yamamoto @RyoYbioinfo
358 Followers 522 Following interping bio models @GoodfireAI · behind EVEE prev; PhD UCLA (Zaitlen / Xiao)
rohan ganesh @certifiedrohan
491 Followers 866 Following obvious ventures | investor: @amilabs @inceptivecom @radicalnumerics @inductive_bio @boltz_bio @ataraxis_ai @syntenyAI and several more
Kartikey Malkani @karmalkani
1 Followers 143 Following
Xuanyi Zhou @Xuanyi_Zhou_
2 Followers 20 Following I'm Xuanyi Zhou. I will be applying to PhD programs for Fall 2027.
Meihao Liao @liaomeihao
2 Followers 230 Following
Karol Muse @karol_muse
41 Followers 624 Following 🔬 Researcher in Phys-Math 🤖 AI Enthusiast & Investor 📈
Eval-Kyaw @Eval_kyaw
2 Followers 67 Following
Rylan Holden @HoldenRylan
3 Followers 78 Following
Unapologetic @mishichoudhary
380 Followers 1K Following Tech Policy, lawyer! Born in India, Aged in US. Digital rights are human rights
winson zhang @winsonzhan83869
0 Followers 14 Following
Guoxia Wang @tian9958
46 Followers 75 Following AI Infra @Baidu PaddlePaddle | Large-Scale Distributed Training & High-Performance Operators | Driving ERNIE LLM efficiency | Creator of FlashMask. 🚀
serious_hamster @grave_hamster
0 Followers 32 Following
Shivam Kumar @Shivam361996
1 Followers 100 Following
Vishrut Malik @MalikVishrut
13 Followers 773 Following
ZPASSER @Justinhk1208
967 Followers 610 Following VPS，GPU使用经验分享，好用工具分享，时不时写点教程，节省你的时间！ 输入改变自己，输出改变世界！ 个人导航站：https://t.co/6RNfNoUEog 整理好用工具和项目，会不断更新。
Andrej Karpathy @karpathy
4.4M Followers 1K Following
Yann LeCun @ylecun
1.3M Followers 800 Following Founder/Chair, AMI Labs; Professor, NYU; Partner, 224 Ventures; Ex-Chief AI Scientist, Meta. Researcher in AI, ML, Robotics, etc. ACM Turing Award Laureate.
Percy Liang @percyliang
127K Followers 424 Following professor of computer science @Stanford @stanfordnlp, co-founder of @togethercompute, https://t.co/7R5THVnJ6u, @simile_ai, pianist
Aran Komatsuzaki @arankomatsuzaki
196K Followers 442 Following Sharing AI research. Early work on AI (GPT-J, scaling, MoE). Ex ML PhD (GT) & Google.
AI at Meta @AIatMeta
869K Followers 367 Following We're building the future of personal superintelligence to help people create, connect, and achieve more in life.
Horace He @cHHillee
55K Followers 620 Following @thinkymachines Formerly @PyTorch "My learning style is Horace twitter threads" - @typedfemale
Google DeepMind @GoogleDeepMind
1.6M Followers 266 Following The engine room of @Google. Building AI safely and responsibly to solve the world’s most complex problems. Join us: https://t.co/jUHQA27iBL
Lucas Beyer (bl16) @giffmana
155K Followers 651 Following Researcher (now: Meta. ex: OpenAI, DeepMind, Brain, RWTH Aachen), Gamer, Hacker, Belgian. Anon feedback: https://t.co/xe2XUqkKit ✗DMs → email
Tim Dettmers @Tim_Dettmers
49K Followers 918 Following Creator of bitsandbytes. Professor @CarnegieMellon and Research Scientist @allen_ai . I blog about deep learning and PhD life at https://t.co/Y78KDJJFE7.
Soumith Chintala @soumithchintala
329K Followers 1K Following Building new things @thinkymachines. Also dabble in robotics at NYU. Cofounded @PyTorch. AI is delicious when it is accessible and open-source.
PyTorch @PyTorch
517K Followers 98 Following Tensors and neural networks in Python with strong hardware acceleration. PyTorch is an open source project at the Linux Foundation. #PyTorchFoundation
Yi Tay @YiTayML
60K Followers 89 Following research scientist @googledeepmind ✨♊, model co-lead/captain of gemini deepthink imo gold medal 🥇, opinions are my own.
Gabriel Peyré @gabrielpeyre
102K Followers 444 Following @CNRS researcher at @ENS_ULM. One tweet a day on computational mathematics.
typedfemale @typedfemale
49K Followers 566 Following a really exciting new account "advanced pytorch user" - @cHHillee alt: @typedalt
Tom Goldstein @tomgoldsteincs
28K Followers 2K Following Professor at UMD. AI security & privacy, algorithmic bias, foundations of ML. Follow me for commentary on state-of-the-art AI.
Beidi Chen @BeidiChen
17K Followers 412 Following Asst. Prof @CarnegieMellon, @amazon Scholar, Prev: Visiting Researcher @Meta, Postdoc @Stanford, Ph.D. @RiceUniversity, Large-Scale ML, a fan of Dota2.
Jürgen Schmidhuber @SchmidhuberAI
224K Followers 0 Following OG of: P and T in ChatGPT, recursive self-improvement, neural distillation, GAN/World Model, deepest learning. Co-authored most-cited AI paper of 20th century
Ross Wightman @wightmanr
25K Followers 1K Following AI / Computer Vision. Ex 🤗. Ex head of Software, Firmware Engineering at a Canadian 🦄. Building AI systems or investing in startups that do it better.
Lachy Groom @lachygroom
301K Followers 922 Following it's pronounced locky. [email protected]. π.com
Zhaoran Wang @zhaoran_wang
6K Followers 1K Following Eng @Nvidia | Prof @NorthwesternU | PhD @Princeton | Reinforcement Learning
Raviraj Jain @ravirajjain
3K Followers 1K Following Partner @ Lightspeed. Focus on Enterprise AI and Frontier tech
Kevin Patrick Murphy @sirbayes
76K Followers 773 Following Research Scientist at Google DeepMind. Interested in Bayesian Machine Learning.
Max Welling @wellingmax
43K Followers 477 Following
Sergey Levine @svlevine
140K Followers 144 Following Associate Professor at UC Berkeley Co-founder, Physical Intelligence
Moritz Stephan @moritz_stephan
4K Followers 359 Following building @hone | prev: @cognition_labs, @StanfordAILab
Nadav Timor @NadavTimor
1K Followers 8K Following AI inference, speculative decoding, open source. Built novel decoding algorithms – default in Hugging Face Transformers (160+ ⭐). Making AI faster + cheaper
Kai Xiao @KaiKaiXiao
923 Followers 151 Following @OpenAI. PhD @MIT, advised by @aleks_madry. Reliable Machine Learning.
kabir @KabirGoel
3K Followers 2K Following taking down prod @cartesia • wasted your tax dollars @cal
Samir Menon @blintzbase
669 Followers 323 Following inference, systems, security, etc @sailresearchco https://t.co/G8yCdm0C2U
Oliver Sieberling @osieberling
979 Followers 796 Following PhD student @MIT, pretraining large neural networks, rnn enjoyer
Michael Y. Li @michaelyli_
3K Followers 530 Following CS PhD @StanfordAILab @StanfordNLP advised by @noahdgoodman and Emily Fox. Prev: undergrad @princeton
Yifan Zhang @yifanzhang_
23K Followers 4K Following PhD at @Princeton University, Princeton AI Lab Fellow. Pretraining Science and RL Science, Prev @NVIDIA @ByteDanceSeed_ @Tsinghua_Uni
noahamsel @noahamsel
233 Followers 84 Following
Xinle Cheng @XinleC295
341 Followers 204 Following PhD @PrincetonCS, previously BS @pku1898 , Generative AI & MLsys
Yu Zhang 🐙🌘 @yzhang_cs
8K Followers 948 Following @Kimi_Moonshot; working on efficient methods for LLMs; disciple of parallel programming; INTP; 🐈
Jean-Gabriel BARTHELE... @JG_Barthelemy
84 Followers 82 Following ML Engineer - Head of Foundation LLM Team at @OVHaiLLM - Training SOTA LLM with lastest research papers
maharshi @maharshii
43K Followers 1K Following learning deeply about life one gradient step at a time - ml perf optimizer @fal
Jerry Tworek @MillionInt
43K Followers 1K Following CEO and co-founder of @coreauto former VP of RL @ OpenAI : reasoning models, o3, o1, GPT4, ChatGPT, Codex, RL for robots cautious AI optimist
Stuart Sul @stuart_sul
4K Followers 168 Following training grok @SpaceXAI, cs phd @Stanford, @HazyResearch
Adam Zweiger @AdamZweiger
3K Followers 638 Following Working on learning @thinkymachines | prev: @cognition @MIT_CSAIL
Schmidt Sciences @schmidtsciences
2K Followers 11 Following Supporting people, projects, and tools to accelerate positive global impact through science and technology.
Liane Galanti @lianegalanti
1K Followers 233 Following PhD student @PrincetonPLI @togethercompute 🪷 ai 🤖
Cathy Di @itsCathyDi
6K Followers 775 Following Co-Founder & CEO @dedaluslabs · prev @PrincetonCS @ptonaialignment · we build wings so your agents can fly 🪽
Junxiong Wang @_junxiong_wang
161 Followers 36 Following Research @togethercompute, CS PhD @Cornell Work on large language models
Charlie Marsh @charliermarsh
52K Followers 1K Following @OpenAI. Building Ruff, uv, ty, and other high-performance Python tools with the @astral_sh team.
John Langford @JohnCLangford
11K Followers 53 Following Solving Machine Learning at Microsoft in New York. https://t.co/ZpdQV4IsHY pandemic past president. https://t.co/MkluiHpWF7 makes RL real. https://t.co/wK8xQaQGwf for thinking out loud.
Misha Goin @mgoin_
3K Followers 484 Following maintainer @vllm_project, inference perf @RedHat_AI (acq @neuralmagic)
Miles Brundage @Miles_Brundage
79K Followers 14K Following AI policy researcher, @lfschiavo wife guy, Executive Director of @AVERIorg, Substacker, fan of animals and sci-fi, views my own
Davis Blalock @davisblalock
15K Followers 171 Following Research scientist @GoogleDeepMind. Past: @Databricks, first hire @MosaicML, @MIT PhD. I post about AI papers + sometimes the business side.
Wentao Guo @WentaoGuo7
1K Followers 202 Following CS PhD student @PrincetonCS, Currently Intern @nvidia, Previously CS MEng + BS @CornellCIS
𝔊𝔴𝔢𝔯𝔫 @gwern
107K Followers 115 Following Internet besserwisser; pedantic, mean reply guy. 𝘞𝘢𝘵𝘢𝘴𝘩𝘪 𝘬𝘪𝘯𝘪𝘯𝘢𝘳𝘪𝘮𝘢𝘴𝘶! (Follow requests ignored due to terrible UI.)
mobicham @mobicham
734 Followers 125 Following I like to shrink dem models 🤏 ML/AI Perf @dropbox Prev. Co-Founder & Principal Scientist @mobius_labs (acquired by @dropbox) PhD @inria
tender @tenderizzation
6K Followers 764 Following hardstuck "senior meme engineer" assembling their promo packet
Aurko Roy @aurko79
2K Followers 446 Following Math & computer science | @AIatMeta (2025-2025) | @GoogleDeepmind (2023-2025) | @GoogleAI (Brain) (2017-2023) | CS PhD @Georgiatech | CS @IITKanpur
NVIDIA @nvidia
2.7M Followers 47 Following The official handle for NVIDIA. Blog: https://t.co/JAn5eKOTBT Support: https://t.co/6ln5FVnA2o All our social media: https://t.co/Uc56dL57Dh
Deepak Pathak @deepakpathak
31K Followers 423 Following Co-Founder & CEO @SkildAI, Faculty @CarnegieMellon. PhD @UCBerkeley; BTech @IITKanpur I study topics in AI (robotics, machine learning & computer vision).You might like
