Rupesh Srivastava @rupspace
Fully open LLM frontiers @MBZUAI IFM Silicon Valley. Previously (co)developed Highway Networks, Upside-Down RL, Bayesian Flow Networks, EvoTorch. rupeshks.cc Santa Cruz, CA Joined September 2014-
Tweets1K
-
Followers3K
-
Following769
-
Likes2K
How can we train small agentic models that are highly capable of terminal use and coding? Announcing OpenThoughts-Agent + OpenThinkerAgent-32B, the strongest Qwen-3 based open-data agentic model: 44.8% avg across 7 agentic benchmarks! (1/n)
@j_foerst Congrats! This is gonna be awesome.
@stochasticchasm Hope, huh? Must be nice 🙂
1/3 Most language models generate text the way a typewriter works. They go left to right, one token at a time. Diffusion language models generate entire sequences by simultaneously refining noise into meaning.
Love it when Jürgen puts things in perspective! 🙂
Tera IPOs coming! $1T sounds like a lot. But $1T is just a 7-m-wide gold cube, thanks to massive inflation since 1971 when $ and gold decoupled. A little house full of gold. To put things in perspective: the 2017 neutron star merger GW170817 produced several earth masses of gold.
Frontier LLMs are converging on efficient, adaptive reasoning. Opus 4.7 lets the model decide how deeply to reason. GPT-5.5 achieves strong results with fewer reasoning tokens. We study a related but more structural question: what 𝗸𝗶𝗻𝗱 𝗼𝗳 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 should we adapt? Last year in SiRA (upper figure), we showed that simulative reasoning (System II), which uses a 𝘄𝗼𝗿𝗹𝗱 𝗺𝗼𝗱𝗲𝗹 to evaluate consequences of actions, yields up to 124% improvement over reactive baselines (System I), and that strong reasoning models (o1, o3-mini) fail as planners without this structure. In our new paper SR²AM (lower figure), we add a learned 𝗰𝗼𝗻𝗳𝗶𝗴𝘂𝗿𝗮𝘁𝗼𝗿 (System III) that self-regulates when to simulate, how far ahead, and when to skip planning entirely. Efficient reasoning is not just shorter reasoning: it is better allocation of simulation.
Thrilled to share that we founded Recursive to create AI that safely conducts experiments on how to improve itself in an open-ended process of endless, automated scientific discovery. As I wrote in my 2019 AI-generating algorithms paper, this will likely be the fastest path to superintelligence. Our work since has shown the power of this approach. Excited to scale up and improve upon ideas like the Darwin Gödel Machine, HyperAgents, ADAS, OMNI, ALMA, The AI Scientist, PromptBreeder, Rainbow Teaming, Automated Capability Discovery, and other work on open-ended and AI-generating algorithms. We’ve assembled a dream team of researchers and significant resources to pursue this vision. My amazing co-founders are pictured here, and we have an all-star team of founding members (we’re over 25 and growing). Please join us if you are interested! Follow our progress @Recursive_SI
Did he just ... wow @fredagainagain1 thank you so much! youtube.com/watch?v=GiXKuk…
Yes!
@charuman wasn't meant as sarcasm it's always nice to see a lab so confident/secure in their capabilities that they can openly publish all their struggles
In this paper, we ask: 𝘏𝘰𝘸 𝘤𝘢𝘯 𝘸𝘦 𝘤𝘭𝘶𝘮𝘴𝘪𝘭𝘺 𝘳𝘦𝘧𝘰𝘳𝘮𝘶𝘭𝘢𝘵𝘦 𝘵𝘩𝘦 𝘤𝘢𝘱𝘢𝘣𝘪𝘭𝘪𝘵𝘺 𝘸𝘦 𝘪𝘮𝘱𝘭𝘦𝘮𝘦𝘯𝘵𝘦𝘥 𝘪𝘯 𝘵𝘩𝘦 𝘧𝘰𝘳𝘮 𝘰𝘧 𝘢 𝘲𝘶𝘦𝘴𝘵𝘪𝘰𝘯?
@finbarrtimbers I think this is likely a difference of scale mainly. If there's enough filtered data to train on, then use that. If there's limited data, train on all.
🍫 CocoaBench v1.0 is out! CocoaBench is a benchmark for unified digital agents, built around open-world tasks that require composing 💻 coding, 👀 vision, 🌐 search. Since our first research preview last December, we have expanded the benchmark substantially with community contributed tasks, and spent months testing and refining the tasks, evaluations, and agent runs. Some takeaways: • Even the best agent system reaches only 45.1% on CocoaBench v1.0. • Coding agents like Codex are already surprisingly strong on general tasks beyond software engineering. • Stronger agents tend to push more of the work into code. • Open source models still lag behind leading frontier models on these general tasks. 👇More on the website and in the paper #AI #Agents #LLM #Benchmark #CocoaBench
🍫 CocoaBench is calling for contributions from the community! Join us and help shape how next-generation agents are evaluated and built🚀✨ #LLM #AI #Agent #CocoaBench More details in the threads 👇
A visually convincing rollout is not the same thing as a useful world model. WR-Arena is built to test the harder question: can a model simulate futures well enough to support action, planning, and reasoning? That’s the shift from simple next-state prediction to realistic world simulation grounded in real-world utility. Paper + code are live. t.co/waRc0MJmwP t.co/ZzN76nOwoI #AI #WorldModels #Benchmarking #EmbodiedIntelligence #PhysicalAI #MachineLearning
@Grad62304977 @kalomaze All networks are mixtures of experts, just gated at unit level :) arxiv.org/abs/1410.1165
The Harbor registry is getting an upgrade. Now, anyone can publish to the registry to make their dataset available to every Harbor user:
Back in beautiful New Haven this weekend for YHack. We’ll be there with K2 Think V2, a fully open-source reasoning system. Hackers! Dig into how it works: huggingface.co/LLM360/K2-Thin…
Yes and no. Very often it turns out that what you think solves the problem is not what actually solves it, and this you only find out by not moving on, but making sure you have experiments that back up the *exact* statement you make removing all reasonable confounders. And that, you get from one of: - public review - extremely strict colleagues - insane self discipline
Delip Rao e/σ @deliprao
69K Followers 5K Following Busy inventing the shipwreck. @Penn. Past: @johnshopkins, @UCSC, @Amazon, @Twitter ||Art: #NLProc, Vision, Speech, #DeepLearning || Life: 道元, improv, running 🌈
Jim Fan @DrJimFan
479K 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.
Soumith Chintala @soumithchintala
310K 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.
Alfredo Canziani @alfcnz
139K Followers 306 Following Musician, math lover, cook, dancer, 🏳️🌈, and an ass prof of Computer Science at New York University
Kyunghyun Cho @kchonyc
86K Followers 2K Following a mediocre combination of a mediocre scientist and a mediocre advisor at @nyuniversity (@CILVRatNYU)
Julian Togelius @togelius
23K Followers 1K Following Researcher. AI, games, markets, open-endedness, evolution. Professor @nyuniversity @NYUGameLab Head of AI @the_nof1 Co-founded @modl_ai Rogueliker.
Miles Brundage @Miles_Brundage
73K Followers 13K Following AI policy researcher, @lfschiavo wife guy, fan of animals and sci-fi, executive director of AVERI (https://t.co/qq9xcmKQas), Substacker, views my own
Sander Dieleman @sedielem
69K Followers 2K Following Research Scientist at Google DeepMind (WaveNet, Nano Banana, Gemini Omni). I tweet about ML, music, generative models (personal account).
Gary Marcus @GaryMarcus
230K Followers 7K Following OG GenAI Skeptic; spoke at US Senate. Warned about hallucinations in 2001. Advocating world models & neurosymbolic AI ever since. Author, Marcus on AI & 6 books
Riley Goodside @goodside
216K Followers 4K Following Mostly screenshots of chatbots since 2022. Formerly: Google DeepMind, Scale.
Jeff Clune @jeffclune
33K Followers 442 Following Co-founder, Recursive. Professor, CS, U. British Columbia. CIFAR AI Chair, Vector Institute. | ML, AI, deep RL, deep learning, AI-Generating Algorithms (AI-GAs)
Nathan Benaich @nathanbenaich
71K Followers 35K Following solo member of superinvestment staff @airstreet @airstreetpress @stateofai @raais
Andreas Kirsch 🇺�... @BlackHC
17K Followers 7K Following My opinions only here. 👨🔬 RS DeepMind 1.8y, Midjourney 1y 🧑🎓 DPhil AIMS 4.5y 🧙♂️ RE DeepMind 1y 📺 SWE Google 3y 🎓 TUM 👤 @nwspk
Sepp Hochreiter @HochreiterSepp
15K Followers 372 Following Pioneer of Deep Learning and known for vanishing gradient and the LSTM.
Felix Hill @FelixHill84
12K Followers 739 Following Research Scientist, Deepmind I try to think hard about everything I tweet, esp on 90s football and 80s music None of my opinions are really someone else's
Shubhendu Trivedi @_onionesque
10K Followers 898 Following Cultivated Abandon. Twitter interests: Machine learning research, applied mathematics, mathematical miscellany, ML for physics/chemistry, books.
Michael Bronstein @mmbronstein
58K Followers 8K Following #DeepMind Professor of #AI @UniofOxford / Director #AITHYRA / Chief Scientist @proximabio / https://t.co/kZpGpDAw4t (opinions are mine) 🤖🧪🧬🎶🐎
rohan anil @_arohan_
43K Followers 2K Following member of technical staff & co-founder of @coreautoai - and continuing to aspire to understand deep learning.
Dileep George @dileeplearning
16K Followers 1K Following Head of AI @AsteraInstitute Prev: AGI @DeepMind, cofounder @vicariousai (acqd by Alphabet), cofounder @Numenta. IIT-Bombay, MS&PhD Stanford. https://t.co/IlsczdBtZo
Rahul @rahul_narava
101 Followers 570 Following RL Community Lead @Cohere_Labs, Pursuing PhD in Reinforcement Learning
clairre17 @LisaAmidon12
6 Followers 461 Following soft like morning light, chaotic like my timeline 🌅 follow back
GalacticMarch @goandstudy117
18 Followers 627 Following "We were never meant to stay planetary. Believer in small and dense intelligence."
Apoorv Khanna @token_wala
182 Followers 725 Following Building AI-native products. High-variance commentary (mostly unhinged). BITS • IIMA • Utah alum
Benhao Huang ✈️ I... @huskydogewoof
1K Followers 787 Following Attracted by Loop Models➰and World Models 🌎 | M.S. student @mldcmu, Prev. @sjtu1896 | Opinions approved by my puppy.
Mingkai Deng @mdeng34
767 Followers 341 Following PhD student @LTIatCMU | MSML @MLDCMU | BA Math-Stats + CS @Columbia | World models and agent models | @IFM_MBZUAI @MSFTResearch
Apoorv Reddy @ApoorvReddy3
3 Followers 703 Following
Malachite @crystalnerd18
420 Followers 948 Following Create AI arts that are somehow inspired by crystals ~
Gaurav @GauravShajepal
359 Followers 3K Following
Louis @Louis9687221579
99 Followers 4K Following Mainline Economics | Idea page | ramblings of a schizo
Pinaki Roychoudhury @fsja_ajau_13145
0 Followers 4K Following
Arsenic @cryoarsenic
0 Followers 80 Following
p13rr0m @p13rr0m
0 Followers 305 Following
Ibrahim Souleymane Mo... @MohamedIbr82078
1 Followers 81 Following
Husain @hmhm1190
13 Followers 583 Following
Richard Zhuang @RichardZ412
1K Followers 808 Following CS @Stanford |Prev. @UCBerkeley @bespokelabsai | LLM Post-Training, Agents, Collective Intelligence
sritee @Sridhaar96
129 Followers 725 Following Interested in ML and Robotics. Research Engineer @GoogleDeepmind
Xuezhe Ma (Max)@ACL @MaxMa1987
2K Followers 431 Following Research Lead @USC_ISI and Research Assistant Professor @CSatUSC PhD at CMU ML/NLP @LTIatCMU @CarnegieMellon
Vishaal Udandarao @vishaal_urao
1K Followers 1K Following @ELLISforEurope PhD Student @bethgelab; Currently @Apple; Previously @GoogleAI @GoogleDeepMind @Cambridge_Uni @RutgersU @iiitdelhi
Swair Bot @SwairBot
0 Followers 788 Following
Omar Khattab @lateinteraction
36K Followers 3K Following asst professor @MIT CSAIL @nlp_mit. https://t.co/VgyLxl0VZz, https://t.co/ZZaSzaRIOF (@DSPyOSS), GEPA, RLMs, Pedagogical RL
dew @GenericHoneydew
16 Followers 2K Following
timur kharisov @normisdeath
80 Followers 756 Following DL optimisation researcher at day, amateur artist at night
Jungian Kant @JungianKant
15 Followers 167 Following CS PhD @UofR | Continual Learning LLMs In search of the oldest ideas
Chris Glaze @chris_m_glaze
1K Followers 4K Following Principal Research Scientist at @SnorkelAI. PhD in computational neuroscience. Previously: @penn @UofMaryland
Bhavishya Sharma @FutureIO
4 Followers 81 Following
Anmol Mekala @anmol_mekala
55 Followers 156 Following applied AI @ Salient | LLM unlearning & benchmarking research | CS @umassamherst, @iitbombay
Croni aka crocromigno... @Cr0cr07
214 Followers 794 Following La Science, la bagarre et la gauche @mjcf_rhone
yati raj @_yatiraj
9 Followers 389 Following remember our ancestors and become good ancestors ourselves
Neham Jain @neham_jain
84 Followers 452 Following 3D everything @MeshyAI | Ex Research @Meta @AdobeResearch @CMU_Robotics
Peter @pedropibil
54 Followers 352 Following
MikeyG @MikeGRighteous
290 Followers 4K Following A prodigy by ghetto standards. Distributed systems wizard. Hedonic treadmill set to MAX.🚀 راحت عليك يا حلو . خبير في كل شيئ
𝕝𝕖𝕠𝕟 @typesuser
262 Followers 5K Following all are rational in the eyes of god ~ foundation models @sistemalabs
Zeeshan @Zshan_ashraf
35 Followers 973 Following
Delip Rao e/σ @deliprao
69K Followers 5K Following Busy inventing the shipwreck. @Penn. Past: @johnshopkins, @UCSC, @Amazon, @Twitter ||Art: #NLProc, Vision, Speech, #DeepLearning || Life: 道元, improv, running 🌈
(((ل()(ل() 'yoav)))... @yoavgo
84K Followers 2K Following
Jim Fan @DrJimFan
479K 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.
Jürgen Schmidhuber @SchmidhuberAI
205K Followers 0 Following OG of: P and T in ChatGPT, 100x deeper learning, meta learning and RSI, neural distillation, GAN/World Model... Co-authored most-cited AI paper of 20th century
Google DeepMind @GoogleDeepMind
1.5M Followers 278 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
Soumith Chintala @soumithchintala
310K 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.
Paul Graham @paulg
3.8M Followers 795 Following
Aran Komatsuzaki @arankomatsuzaki
182K Followers 375 Following Sharing AI research. Early work on AI (GPT-J, scaling, MoE). Ex ML PhD (GT) & Google.
Eric Jang @ericjang11
135K Followers 4K Following
Alfredo Canziani @alfcnz
139K Followers 306 Following Musician, math lover, cook, dancer, 🏳️🌈, and an ass prof of Computer Science at New York University
Edward Grefenstette @egrefen
46K Followers 916 Following FR/US/GB AI/ML Person, Director of Research at @GoogleDeepMind, Honorary Professor at @BOLD_Lab_AI, @ELLISforEurope Fellow. All posts are personal.
Kyunghyun Cho @kchonyc
86K Followers 2K Following a mediocre combination of a mediocre scientist and a mediocre advisor at @nyuniversity (@CILVRatNYU)
Thomas G. Dietterich @tdietterich
62K Followers 651 Following University Distinguished Professor (Emeritus), Oregon State Univ.; Former President, AAAI; Currently Chair CS Section of ArXiv
Peyman Milanfar @docmilanfar
113K Followers 578 Following Distinguished Scientist at Google. National Academy of Engineering. Computational Imaging ∩ AI. Posts are personal opinions
Julian Togelius @togelius
23K Followers 1K Following Researcher. AI, games, markets, open-endedness, evolution. Professor @nyuniversity @NYUGameLab Head of AI @the_nof1 Co-founded @modl_ai Rogueliker.
Benhao Huang ✈️ I... @huskydogewoof
1K Followers 787 Following Attracted by Loop Models➰and World Models 🌎 | M.S. student @mldcmu, Prev. @sjtu1896 | Opinions approved by my puppy.
Mingkai Deng @mdeng34
767 Followers 341 Following PhD student @LTIatCMU | MSML @MLDCMU | BA Math-Stats + CS @Columbia | World models and agent models | @IFM_MBZUAI @MSFTResearch
opentraces @opentraces
59 Followers 4 Following What the agent sees, does, and changes. A local-first evidence layer to unlock trapped agent traces for safe sharing via @huggingface
ClaudeDevs @ClaudeDevs
541K Followers 2 Following Official updates for developers building with @ClaudeAI
Institute of Foundati... @IFM_MBZUAI
284 Followers 8 Following
Mario Zechner @badlogicgames
54K Followers 1K Following Armin's handler at https://t.co/B05ybKGkzx. Old man yelling at Claudes. https://t.co/Q1wG57v1yc https://t.co/mnOoWUr0TO https://t.co/8i5vIRE0Wn
Richard Zhuang @RichardZ412
1K Followers 808 Following CS @Stanford |Prev. @UCBerkeley @bespokelabsai | LLM Post-Training, Agents, Collective Intelligence
Massimo Caccia @MassCaccia
2K Followers 668 Following Post-Training @Cohere 🇨🇦 Formerly @ServiceNowRSRCH, @Mila_Quebec, @GoogleDeepmind, @AWScloud, @SpotifyResearch.
John Yang @jyangballin
6K Followers 1K Following CS PhD @Stanford. Created @SWEbench (multi-lingual/modal); SWE-agent; SWE-smith; InterCode; CodeClash; ProgramBench
Clive Chan @itsclivetime
28K Followers 3K Following perplexity per picojoule @anthropicai // prev jalapeno @openai, dojo @tesla
Damek @damekdavis
8K Followers 1K Following Prof @Wharton stats / optimization & ML / AI for math / benchmark https://t.co/a1OxJdthkM / course https://t.co/bfOIEx0lHj
Omar Khattab @lateinteraction
36K Followers 3K Following asst professor @MIT CSAIL @nlp_mit. https://t.co/VgyLxl0VZz, https://t.co/ZZaSzaRIOF (@DSPyOSS), GEPA, RLMs, Pedagogical RL
Harbor Framework @harborframework
1K Followers 3 Following
Ryan Marten @ryanmart3n
2K Followers 2K Following Building @harborframework and @terminalbench with @alexgshaw
stochasm @stochasticchasm
7K Followers 2K Following pretraining lead @arcee_ai • 25 • opinions my own
Bowen Tan @BowenTan8
166 Followers 203 Following PhD student @LTIatCMU @SCSatCMU; Member @llm360; Student researcher @Google
Ariel @redtachyon
5K Followers 277 Following p/hd | Big RL energy | RS @ fruit company (not speaking for the company though) | Prev. {Meta FAIR; Gym(nasium)} | Technology Sibling
Kanishk Gandhi @gandhikanishk
2K Followers 1K Following Phd CS@Stanford @StanfordNLP, Computation and Cognition; w/ Noah Goodman | Prev: @MSFTResearch @LakeBrenden @NYUDataScience, @IITKanpur, @Path_AI
LMSYS Org @lmsysorg
16K Followers 199 Following Large Model Systems Organization: Join our Slack: https://t.co/vzYOTP4w6C. We developed SGLang https://t.co/OjwQadINKU, Chatbot Arena (now @arena), and Vicuna!
Sangyun Lee @sang_yun_lee
1K Followers 467 Following PhD student @CMU_ECE | ex-intern @nvidia | Generative models
Sungjin Ahn @SungjinAhn_
3K Followers 1K Following Prof@KAIST, Chief Dreamer of Machine Learning & Mind Lab https://t.co/ato9yodtm5
Arif Ahmad ✈️ CVP... @arif_ahmad_py
862 Followers 7K Following We are in the world model era now. Currently IFM. Prev. @GoogleDeepMind and @Nvidia
Catherine Olsson @catherineols
22K Followers 1K Following Hanging out with Claude, improving its behavior, and building tools to support that @AnthropicAI 😁 prev: @open_phil @googlebrain @openai (@microcovid)
Zed @zeddotdev
80K Followers 84 Following A next-generation code editor that enables high-performance collaboration with AI and your team. https://t.co/4Ua0UqLrsv
Zhiqing Sun @EdwardSun0909
20K Followers 1K Following Lead agent research @Meta MSL TBD Lab. previously posttraining/agent research @OpenAI. CS PhD @LTIatCMU
Samuel Albanie 🇬�... @SamuelAlbanie
8K Followers 1K Following gemini evals & post-training @GoogleDeepMind
AI-Driven Research fo... @ai4research_ucb
1K Followers 9 Following
Yifan Zhang @yifanzhang_
14K Followers 3K Following PhD at @Princeton University, Princeton AI Lab Fellow. RL & LLM Reasoning, Pretraining & Language Modeling. Prev @ Seed @Tsinghua_Uni
Ellie Cheng @ellieyhc
351 Followers 262 Following PhD student @mit_csail | Programming systems for ML/AI | don’t ask how many cups of coffee I’ve drank
InclusionAI @TheInclusionAI
2K Followers 19 Following AI Lab @AntGroup, we envision AGI as humanity's shared milestone. Our Language Model @AntLingAGI and LLaDA, Embodied AI @robbyant_brain, OSS projects AReaL etc.
Ant Ling @AntLingAGI
10K Followers 5 Following MoE model series with foundation (Ling), reasoning (Ring) and any-to-any (Ming) from Ant Group’s AGI initiative, @TheInclusionAI. https://t.co/6LEkFloA1Y
Saining Xie @sainingxie
40K Followers 2K Following cofounder & chief science officer at @amilabs | faculty @nyu_courant | prev: @googledeepmind @meta (fair) @ucsandiego | ynwa
Stella Li ✈️ ICML... @StellaLisy
4K Followers 540 Following PhD student @uwnlp | visiting researcher @AIatMeta | undergrad @jhuclsp #NLProc
Kazuki Irie @kzkirie
395 Followers 188 Following Computer scientist @Yale @WuTsaiYale, advised by @HandsomeDanYale. KI, der KI erschafft. All tweets are wrong, but some are useful.
Xuezhe Ma (Max)@ACL @MaxMa1987
2K Followers 431 Following Research Lead @USC_ISI and Research Assistant Professor @CSatUSC PhD at CMU ML/NLP @LTIatCMU @CarnegieMellon
Syrine @syrineblk
151 Followers 449 Following Senior Researcher at Microsoft| Stanford Data Science Fellow| PhD WSU CS | IBM PhD Fellow 2021 | MIT EECS Rising Star https://t.co/p6aZaJSV34
Xingzhi Sun @XingzhiSun
34 Followers 413 Following CS PhD student @Yale (@KrishnaswamyLab). Research Intern @AIatMeta. Prev Research Intern @Genentech Regev Lab. AI4Science | LLM | GenAI | Single cell dynamics
Hector Liu @waterluffy
324 Followers 327 Following Building Institute of Foundation Models (https://t.co/eSDH1CD8ly) Views Mine. @LLM360, LLM, (del)NLP, Computational Linguistic(del)
Prithviraj (Raj) Amma... @rajammanabrolu
8K Followers 681 Following Reinforcement Learning and Language. Assistant Prof @UCSanDiego. Research Scientist @Nvidia.
Makesh Narsimhan @MakeshNarsimhan
271 Followers 492 Following NLP Research @NvidiaAI | Prev @AmazonScience, CS @UWMadison | Eat Right, Stay Fit, Die Anyway.

































