dineshreddy

✨ 𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗩𝗟𝗠 𝗥𝘂𝗻 𝗢𝗿𝗶𝗼𝗻 – our new visual agent that 𝘴𝘦𝘦𝘴, 𝘳𝘦𝘢𝘴𝘰𝘯𝘴, 𝘢𝘯𝘥 𝘢𝘤𝘵𝘴. 👉 chat.vlm.run Today’s frontier Companies – ChatGPT, Claude, Gemini – can describe what they see … But they can’t 𝗮𝗰𝘁 on it.

@LerrelPinto @Meta Congratulations @LerrelPinto

ARI is joining @Meta! Over the past year, we have been building ARI (Assured Robot Intelligence) with the mission to build industry-grade physical AI for humanoids. The ARI stack is built on human experience, condensed into actionable tokens that can be rapidly adapted to real-world hardware. But the most rewarding part of ARI has been the people. I feel truly blessed to have worked alongside some of the world's best roboticists, a top-notch investor pool led by @aixventureshq, and the many supporters pushing for us behind the scenes. Starting next week, ARI will join the Meta Superintelligence Labs (MSL) to continue advancing frontier robotics models that advance personal superintelligence in the physical world. We have the potential to transform AI that can think and talk to AI that can do, assisting humans safely and reliably in the physical world. To the many people behind the scenes who supported us: Thank you! This is just the beginning. More in the Bloomberg article:

15 years with NVIDIA GPUs — and now I’m at NVIDIA GTC. It started in undergrad experimenting with GPUs while playing video games. In 2015, I trained my first autonomous driving model. During my PhD at CMU, GPUs powered many of the models I trained, and later I deployed ML systems into production at Amazon and now at VLM Run. Being at NVIDIA GTC feels like a full-circle moment. What’s exciting is seeing how the technology has evolved. GPUs that once powered gaming and research experiments are now driving agents, robotics, and physical AI at scale — the same areas I’ve been working on. Walking around GTC, meeting incredible people from NVIDIA and across the ecosystem, seeing robots in action, and even spotting futuristic robotics — it’s inspiring to see how far things have come. Excited for what’s next 🚀 If you're working on visual agents or physical AI, I’d love to connect. And if you're interested in joining us, we’re hiring. #NVIDIAGTC #AI #PhysicalAI #Agents #Robotics #GPUs

What if visual AI agents could help give feedback on exercise? @jeremyparkphd recently reviewed Orion visual AI agent for providing deadlift feedback. He raises the question: what if visual intelligence could be made accessible for applications in exercise, all through a chat interface? Read the Substack blog here: jeremyparkphd.substack.com/p/visual-agent…

@GeminiApp The bigger question is how do we change our exam format for SATs and JEE in this ai world which can bring out the best in these young minds.

Building on last week’s announcement of practice SATs in Gemini, you can now also take full-length mock JEE Main tests in Gemini at no cost.

@elonmusk great to see FSD improvements but i am seeing that the auto steer degrade. may be we can use some of the stack from FSD in autosteer as well

Tesla will stop selling FSD after Feb 14. FSD will only be available as a monthly subscription thereafter.

Artifacts will be the Internet like for Multi Modal AI Agents

Introducing Artifacts for Orion 🧵 vlm.run/blog/introduci… Multi-modal outputs as first-class objects. No URL juggling. No storage glue. Generate images, videos, masks, and 3D reconstructions – and get typed references you can store, reuse, and compose.

@JeffDean @Waymo Really impressive how far autonomous driving has come from research to production.

My first @Waymo ride on the highway as I head from a meeting in Redwood City to Gradient Canopy in Mountain View. Nice work, Waymo team! 🎉

@vkhosla I agree as well

Very likely

@bindureddy I do not think anthropic will be sold to google.

Anthropic shouldn’t sell to Google or anyone else They have excellent models that are on the verge of AGI and will be enormously successful as an independent entity!

@menhguin For visual tasks Gemini is one of the best we extensively route some tasks to Gemini at @vlmrun. You can call me a fanboy for Gemini models.

I wonder what a Gemini superfan is like. I've seen superfans for ChatGPT, Claude, Grok, Deepseek, Kimi, but I've never seen someone get really hyped about Gemini models.

We've been hard at work at @vlmrun to bring you the best experiences. Wanted to share how 2025 has been a crazy year for model usage: I alone used 1.5 billion tokens for coding, while our customers crossed trillions. Excited to end the year with this milestone as our customers move toward quadrillions of tokens in the new year. Wishing everyone a happy new year!

We’re hiring at @vlmrun We launched Orion just a few weeks ago. It's already going into production with enterprise users, with more deployments lined up for Q1 2026. Demand is outpacing our team size, so we are scaling up. If you want to build the runtime + interface for visual intelligence (docs, images, video) – DM me or checkout app.dover.com/jobs/vlm-run.

@docmilanfar Shouldn’t the poster have a minimum size at conferences?

worst poster award

Some other cools things at the conference. tesla Robot and the gemini team giving a talk.

We are hiring: - ML Engineer/ Researchers - Interns I generated this image using VLM Run Agents. I created several images from a single prompt, picked the best one, and then refined it further to get this final result. chat.vlm.run/c/436df246-8da…

📍 Live from #NeurIPS! We're looking for the best minds to join us at @vlmrun. If you're passionate about building the future of Vision Language Agents, come say hi or slide into our DMs. We are hiring: app.dover.com/apply/vlm-run/… #AI #MachineLearning #Hiring #VLM #GenAI

@CreatedByJannn You can do the exact same thing in a chat with @vlmrun. Showing results for text -> image(nano banana) -> object segmentation -> 3d reconstruction(using orion visual agent) Check the results out at: chat.vlm.run/c/3fcd6b33-266… try for free at chat.vlm.run

@BamChrisAI You can do the exact same thing in a chat with @vlmrun . Showing results for text -> image(nano banana) -> object segmentation -> 3d reconstruction(using orion visual agent) Check the results out at: chat.vlm.run/c/3fcd6b33-266… try for free at chat.vlm.run.

And now the next one: Engineering drawing -> nano banana to create 3d view -> SAM3D to create 3d model. Not perfect, but a start.