Yash Bhalgat

💡 Should you do a PhD in AI (2025–26)? 🎓 🔗: yashbhalgat.github.io/blog/phd-or-no… Every October, students considering PhD applications ask me: is a PhD still the right path in AI? ⚖️ ⚠️ After a few years moving between academia (@UniofOxford's @Oxford_VGG) and industry (@QCOMResearch, @Meta Reality Labs, and a few startups), I’ve seen both sides of the research world. And the truth is: they’ve never felt further apart. 🌟 Today, most of the *scale-driven* work -- world models, video generation, large VLMs -- happens in industry. Compute access, data scale, and iteration speed make that inevitable. But academia still matters: it’s where new ideas, theory, and deep conceptual work often begin. The difference now is knowing what not to work on. 📢 I’ve written a longer, no-BS post on this -- what makes a PhD worth it, when it doesn’t, and how to think about your timing. 🧭 Read it, share it, debate it -- just don’t decide by inertia. Full post here: yashbhalgat.github.io/blog/phd-or-no… #PhD #AI #ArtificialIntelligence #MachineLearning #PhDLife #Research #AcademicTwitter #GradSchool #CareerAdvice

@venturetwins hi Justine, can I have access? 👀

Truly blown away by a new AI image model launching this week ✨ Finally, you can generate photos that actually look like you! It's so much better than everything I've tried - from LoRAs to NB Pro. Onboarding some early testers. DM or comment if you want access 👀

I grew up watching @lang_lang's performances on youtube. This is so fun... absolutely unreal. 🤩

@NandoDF Just seeing this when trying to access MAI Playground: "This region isn't supported yet. We're working to expand access as soon as possible."

@NandoDF when can we try this out? 👀

Super proud of our small image-gen team that in less than 6 months achieved this amazing result. .... and they're already cooking the next one, but I'm encouraging them all to take a vacation😎

@EvanGruhlkey @Airbound_Aero agreed. There's so many underrated startups in the Indian ecosystem sadly no one knows about.

@ysbhalgat @Airbound_Aero India’s startup scene is quietly producing some serious hardware companies.

Wow, this is super cool! Didn't know about this Indian drone startup before. @Airbound_Aero

Happy Pi Day! ⭕

@dchaplot Congrats @dchaplot! :)

I'm joining SpaceX and xAI, working closely with Elon and team to build superintelligence. Together SpaceX and xAI combine physical and digital intelligence under a leader who understands hardware at the deepest level. Add a high-agency culture with frontier-scale resources, and you get the possibility to achieve something truly unique. I’m excited to advance the fields I’ve obsessed over for years, from robotics research to building AI models on the founding teams of Mistral and TML. Both were extraordinary journeys with extraordinary people that shaped how I think about building intelligence from the ground up. Grateful for everything that brought me here and can’t wait to get started.

really fun to watch the finalists work through the 5 problems youtube.com/watch?v=S8enYu…

so, 2026 = 2^(22/2) - 22 noticed this cute little coincidence while working through Problem 3 from this year’s MIT Integration Bee Finals 😅

@natanielruizg This is really cool work! congrats :))

Excited to show some surprising inventions on generative multiplayer games we made at Google with Stanford. We call the work MultiGen. I've always been inspired by early studios like id Software with Doom or Blizzard with Warcraft bringing networked video games to the next level. We are at the point in history where we can make strides like them, but for generative games. It's a strange feeling to be in the age of generative video games while still discovering how exactly to train the models and design the tools that make them useful. All of the tools that have been invented for classic game engines need to be redesigned for generative games. For example level and world design is not entirely possible with existing technology. We introduce editable memory to diffusion game engines that allow for design of new levels via a minimap. But we can easily imagine how this can be expanded with different creation tools. The end goal of this research direction is to allow game designers to be able to guide the generation process of their world, at the granularity that they prefer. Editable memory also allows us to add multiplayer to Generative Doom. We were amazed when we saw GameNGen some years ago, and now you can play it live with friends in real-time, on your couch or even online. Shared representations like our editable memory seem like the future for this type of experience. Models are, in some cases, expensive and approximate encoders but great interpolators and extrapolators. Leveraging their strengths lets you have completely new experiences that can be realized now and not in the distant future. This work was started at my previous team and continued in collaboration with Stanford. Congratulations to all for the discoveries.

@holynski_ Very cool work, Aleks and @Haian_Jin! :)

zip zap, it's time to map! sota 3d reconstruction, but many times faster than VGGT and pi3.

Just when you thought the streaming wars couldn't get more interesting… 👀 @BenAffleck's AI startup, InterPositive, is joining @netflix. It’s not just a production deal; Netflix is acquiring the technology and the entire team. Video-gen 🤝 Hollywood icon 🤝 Streaming giant

@ChujieZheng not this again lol 😅

Why is everyone leaving @Alibaba_Qwen in the last 2 days?? cc @alibaba_cloud

"just realized"? bruh.

@crude2refined @francoisfleuret log(avg(softmax)) = log(sum(softmax)) - log(N) ?

@francoisfleuret Why log of avg of soft max not an option?

You get N vectors of logits for the same prediction, you got with

@Eng_Hemdi @cerspense Hahaha. There's still a thrill in building things by hand.

@cerspense @ysbhalgat maybe it’s too late to learn Blender 😁

I built an agentic system that taught itself the Blender donut tutorial by watching it on YouTube. It watched the tutorials, extracted the steps, filled in the gaps in own tooling and completed the entire thing autonomously.

@Acerola_t Crazy that this is a rendering from a simulation. Amazing realism 😮

here's an actual real time realistic 3D ocean simulation I wrote a few years ago without AI just by reading some papers. maybe someday Gemini can catch up to 30 year old rendering tech

What a moment though 😎