RT Bilawal Sidhu: Gave google omni a sketched camera path and asked it to generate drone POV footage.
RT Bilawal SidhuGave google omni a sketched camera path and asked it to generate drone POV footage.
X / Twitter 6297 stories
RT Eric Jang: .@tinkerapi by TML is a highly underrated product. It should be part of the “starter post training” infra for any newco
RT Eric Jang.@tinkerapi by TML is a highly underrated product. It should be part of the “starter post training” infra for any newco
RT Sam Sheffer: omni continues to blow my mind (left original / right generated) waymo looks sick in matte black
RT Sam Shefferomni continues to blow my mind(left original / right generated) waymo looks sick in matte black
RT Nathan Lambert: Gemma 4 adoption numbers outpacing Qwen 3.5/3.6 for the same sized models is a big shift in the international balance of influence …
RT Nathan LambertGemma 4 adoption numbers outpacing Qwen 3.5/3.6 for the same sized models is a big shift in the international balance of influence via open…
Forecasting Scientific Progress with Artificial Intelligence https://arxiv.org/abs/2605.22681 Turns out AI is just as bad at forecasting biology and p…
Forecasting Scientific Progress with Artificial Intelligencehttps://arxiv.org/abs/2605.22681Turns out AI is just as bad at forecasting biology and physics breakthroughs as we are. To be fair, most breakthroughs…
Memorial Day. Today we honor those who gave everything to defend our democracy. Democracy isn't guaranteed; it's a precious inheritance that requires …
Memorial Day. Today we honor those who gave everything to defend our democracy. Democracy isn't guaranteed; it's a precious inheritance that requires our constant care. 🇺🇸
RT F.O.L.A: Google Omni might be too powerful 🫥
RT F.O.L.AGoogle Omni might be too powerful 🫥
RT Garry Tan: Thinking Machines is impressive. In a couple hours I just fine tuned my own Qwen3.5-397B model this afternoon. Fast usable multimodal is…
RT Garry TanThinking Machines is impressive. In a couple hours I just fine tuned my own Qwen3.5-397B model this afternoon. Fast usable multimodal is also going…
Added a DeepSeek Sparse Attention (DSA) from-scratch implementation to my LLMs-from-scratch repo thanks to an awesome new reader contrib. With motivat…
Added a DeepSeek Sparse Attention (DSA) from-scratch implementation to my LLMs-from-scratch repo thanks to an awesome new reader contrib. With motivation, overview, and GPT-style model reference…
The new White House policy requiring green card applicants to apply from outside the US is a capricious attack on legal immigration. It will hurt fami…
The new White House policy requiring green card applicants to apply from outside the US is a capricious attack on legal immigration. It will hurt families,…
Harvard University just voted to limit the number of A grades given in undergraduate classes to about 20% of the class. I’m not in favor of this. It …
Harvard University just voted to limit the number of A grades given in undergraduate classes to about 20% of the class. I’m not in favor of…
RT Google DeepMind: Project Genie 🤝 @GoogleMaps Street View You can now take real U.S. places and transform them into new, interactive worlds. 🌍
RT Google DeepMindProject Genie 🤝 @GoogleMaps Street ViewYou can now take real U.S. places and transform them into new, interactive worlds. 🌍
It's been *almost* a bit quiet around LLM architecture releases in the past two weeks 😅 Interesting tidbit is the parallel block design. Via the Cm…
It's been *almost* a bit quiet around LLM architecture releases in the past two weeks 😅Interesting tidbit is the parallel block design. Via the Cmd-A the…
New course: Build AI agents that generate images and videos — an under-explored frontier. A key to performance is having the agent evaluate its own o…
New course: Build AI agents that generate images and videos -- an under-explored frontier. A key to performance is having the agent evaluate its own output,…
Collaborative AI runs on interactivity: machines and people, working in real time, across every modality. Solving it takes a community, join us.
Collaborative AI runs on interactivity: machines and people, working in real time, across every modality. Solving it takes a community, join us.Thinking Machines: We are offering…
We would love to see more collaboration and research in the field of human-AI interactivity. Check it out!
We would love to see more collaboration and research in the field of human-AI interactivity. Check it out!Thinking Machines: We are offering grants of $100,000 +…
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the…
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join…
I only recently read more about the concept of system accidents by Charles Perrow, very insightful and relatable.
I only recently read more about the concept of system accidents by Charles Perrow, very insightful and relatable.
New article: a visual tour of recent LLM architecture advances, from Gemma 4 to DeepSeek V4. I focus on long-context efficiency tweaks like KV sharing…
New article: a visual tour of recent LLM architecture advances, from Gemma 4 to DeepSeek V4.I focus on long-context efficiency tweaks like KV sharing, per-layer embeddings,…
New course: Transformers in Practice. You'll get a practical view of how transformer-based LLMs work, so you can reason about their behavior, diagnose…
New course: Transformers in Practice. You'll get a practical view of how transformer-based LLMs work, so you can reason about their behavior, diagnose problems like slow…
more demos on Interaction Models collaboratively doing system design, reading papers, fact-checking with live generative UI
more demos on Interaction Models collaboratively doing system design, reading papers, fact-checking with live generative UISeongsik Kim: 1. (System design) - The Interaction Models see your…
A little talk on what we can learn from implementing LLM architectures from scratch in Python and PyTorch. And how I approach new open-weight models, …
A little talk on what we can learn from implementing LLM architectures from scratch in Python and PyTorch. And how I approach new open-weight models, compare…
Interesting paper. What I like about this is that it is a relatively low-commitment attention modification. I.e., one can use it during most of traini…
Interesting paper. What I like about this is that it is a relatively low-commitment attention modification.I.e., one can use it during most of training, switch back…
Cluster magicians and GPU whisperers, come join us! We’re looking for supercomputing engineers to build the infrastructure behind real-time interacti…
Cluster magicians and GPU whisperers, come join us!We’re looking for supercomputing engineers to build the infrastructure behind real-time interactive models, Tinker, and large-scale training: scheduling, storage,…