ToddlerBot 2.0 is released🥳! Now Toddy can also do cartwheels🤸! We have added so many features since our first release in February; see github.com/hshi74/toddler… for more details. Threads🧵(1/n)
Continuing the journey of optimal LLM-assisted coding experience. In particular, I find that instead of narrowing in on a perfect one thing my usage is increasingly diversifying across a few workflows that I "stitch up" the pros/cons of:
Personally the bread & butter (~75%?) of…
The thing about RL envs, or RL env libraries, is that envs should be completely independent from the algorithmic details of the learner.
An env shouldn't know anything about vLLM, or about wandb, or about OpenAI clients - it should just implement the internal env logic.
Repetition rewires
Repetition rewires
Repetition rewires
Your brain rewires what you repeat - for better or worse. Neuroplasticity doesn’t know the difference.
Be mindful of what you repeat.
What if we could evolve AI models like organisms in nature, letting them compete, mate, and combine their strengths to produce ever-fitter offspring?
Excited to share our new work: “Competition and Attraction Improve Model Fusion” presented at GECCO’25🦎 where it was a runner-up…
xAI will soon be far beyond any company besides Google, then significantly exceed Google.
Companies in China will be the toughest competitors, because they have so much more electricity than America and are super strong at building hardware.
xAI will soon be far beyond any company besides Google, then significantly exceed Google.
Companies in China will be the toughest competitors, because they have so much more electricity than America and are super strong at building hardware.
The @xai Grok 2.5 model, which was our best model last year, is now open source.
Grok 3 will be made open source in about 6 months.
huggingface.co/xai-org/grok-2
Many years ago, I took CS 231N at Stanford, taught by none other than the legendary @karpathy. I remember one of the feedback for the class highlighted the moments where Andrej said "I don't know" for questions he didn't know. That is the beginning of knowledge.
Many years ago, I took CS 231N at Stanford, taught by none other than the legendary @karpathy. I remember one of the feedback for the class highlighted the moments where Andrej said "I don't know" for questions he didn't know. That is the beginning of knowledge.
Excited to share we got 5 papers accepted to #EMNLP2025! Congrats to all the students! And great thanks to all the collaborators!
1. The Hallucination Tax of Reinforcement Finetuning (arxiv.org/abs/2505.13988). led by @linxins2 and @taiwei_shi
🚀 Excited to share that our paper "𝗥𝗲-𝗔𝗹𝗶𝗴𝗻: 𝗔𝗹𝗶𝗴𝗻𝗶𝗻𝗴 𝗩𝗶𝘀𝗶𝗼𝗻 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝘃𝗶𝗮 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹-𝗔𝘂𝗴𝗺𝗲𝗻𝘁𝗲𝗱 𝗗𝗶𝗿𝗲𝗰𝘁 𝗣𝗿𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻" has been accepted to EMNLP 2025 (Main Track)! 🎉
Large…
If you're fine-tuning LLMs, Gemma 3 is the new 👑 and it's not close. Gemma 3 trounces Qwen/Llama models at every size!
- Gemma 3 4B beats 7B/8B competition
- Gemma 3 27B matches 70B competiton
Vision benchmarks coming soon!
Programming with AI is insanely fun. Process is:
1. generate code
2. read & understand code that was generated
3. make small changes "manually" (still with great autocomplete)
4. test & debug
5. make big changes with new prompt
6. go back to step 1
Pure vibe coding skips step 2…
19 Followers 392 FollowingIIT Bombay EE 2018 भारतीय
अभियंता, Network Security, Red Team, White Hat, Backend developer, Python, Lang-chain, LLM,
Bug Bounty,
DHH, Music production 🎁
231 Followers 3K FollowingGuiding ElonMusk's vision for a better future through SpaceX, Tesla, Neuralink, and more. & Tech enthusiast, dream chaser, and innovation advocate
43K Followers 123 FollowingA research institute & Deemed University for Natural Sciences, Mathematics, Computer Science & Science Education, under the Department of Atomic Energy.
5K Followers 40 FollowingThe Institute of Mathematical Sciences (IMSc) is a national institute for fundamental research in the mathematical and physical sciences
38K Followers 1K FollowingCo-creator of GitHub Copilot, Dropbox Paper, AI Tinkerers, Hackpad, MobileCoin, Minion AI, etc. Working on @PerplexityComet. Survivor 🎗️
4K Followers 2K FollowingResearch Scientist at @Meta Fundamental AI Research (FAIR), New York. Previously: Postdoc @Caltech, PhD @PrincetonCS, Undergrad @Tsinghua_Uni.
751 Followers 203 Followingp/hd | Big RL energy | 0.71 |research⟩ + 0.71 |engineer⟩ @ Meta, but never speaking on behalf of the company | Prev. lead maintainer of Gymnasium
2K Followers 969 FollowingResearch Scientist @ Google Deepmind.
TL on Agents and Project Mariner in Gemini.
Prev: cofounder @AdeptAILabs, Google Brain.
Working on creating AGI.
50K Followers 3K FollowingAI alignment + LLMs at Anthropic. On leave from NYU. Views not employers'. No relation to @s8mb. I think you should join @givingwhatwecan.
38K Followers 991 FollowingCreator of bitsandbytes.Research Scientist @allen_ai and incoming professor @CarnegieMellon. I blog about deep learning and PhD life at https://t.co/Y78KDJJFE7.