Super excited to share that we, Morpheus AI, come out of stealth today and are joining @Roblox.
It was an incredible honor to join Xun as founding member, grow the team, and witness our achievements in all the months that followed.
As next steps, we will continue building our foundational video world models. I’m personally excited most about optimizing everything optimizable!
I'm excited to announce that the Morpheus AI team is joining Roblox!
Over the past two years, I’ve focused on developing the foundational architectures behind modern video world models, including Self Forcing and AR-DiT. This work unlocked something unprecedented: the ability to move beyond offline, pre-rendered AI video generation and instead simulate interactive worlds in real time. Realizing the massive potential of this technology is what drove me to found Morpheus in August 2025. In the months since, our incredible team has pushed those boundaries further than we ever thought possible.
We've always believed video world models will reshape how games are created. Roblox Reality is an ambitious bet on that exact future, and it lines up perfectly with what we set out to do: bridging the gap between deterministic game engines and generative world models. Joining Roblox means our technology will help power experiences that reach millions of players every day.
To our team, to @a16z and other investors, and to the advisors, partners, and supporters who believed in this from the very beginning — thank you.
We're just getting started. Excited to build this at scale.
Check out this new work from our team!
Pixel space, frame-wise, truly “real-time” generation
“Real-time” video models are finally real-time!
Introducing PixVid: lifting a latent video model to generate one frame at a time, in pixel space. No VAE, no temporal compression.
The payoff? Fast high-res generation, precise editing, and ultra-low control-to-screen latency!
33
1,011
Check out @MennDennis’s work on mixed resolution generation!
MPEG Diffusion for video gen: high-resolution on important subjects, low-res on the background to save compute while keeping visual quality.
It features:
1. Near full-res quality
2. DiT self-select high-res regions
3. Up to 4.8× faster (30% tokens)
15
699
Aryan V S retweeted
Introducing FrontierPhysics - a benchmark evaluating agents for end-to-end physics research
Our core team from physics labs at Berkeley, Caltech, and Harvard recruited 42 physicists from 26 institutions. 28 of them published papers on Nature, Nature Physics, PRX, PRA*
1/7 🧵
Aryan V S retweeted
drisspg.github.io/nuggets/Wh…
not really a post - more a fun visual claudagraphic. TLDR pack all fp32 bit patterns into a tensor and see what patterns emerge for `equivalent ` forms of unary ops with fast math set
i want to go back to the time i actively participated in CTFs, and got introduced to schizo problems based on ideas like this one
There is an (unpatchable?) supply-chain zero-day in PyTorch autograd that allows a regularization term to use the model's weights as RAM and SGD as its CPU clock to run any computation during training.
1
11
596
Aryan V S retweeted
Introducing Robo Use: API to physical worlds for agents
One interface, 32 embodiments, 100+ scenarios, and 11 agents and models
We aggregated all academic simulator scenarios and ran 2500+ trials with initial results.
Preview them or try it yourself 🧵
Aryan V S retweeted
Have been busy with work, but finally I find the time to finish the blogpost on @GPU_MODE @coreauto's QR competition that happened 3 months ago.
I didn't win the 1st prize, but my panel QR kernel is definitely the fastest (by a large margin)
gau-nernst.github.io/b200-qr…
Aryan V S retweeted
Super happy to share our intention to join forces with NVIDIA in a $12,930,300,000 acquisition 💛💚
10 years after starting Hugging Face, open-source AI is at an inflection point. Thanks to the community, we’ve shown that it can be a complement, and even an alternative, to closed-source APIs. But for it to happen at larger scale, it needs more compute, more support, more collaboration and more visibility. That’s why we went to talk to Jensen, who offered to do exactly that with us.
In addition to doubling down on NVIDIA’s massive contributions to open-source AI (I called them the “King of American open-source AI” earlier this year), they’ve committed to strongly supporting Hugging Face and our mission while keeping the platform open, independent and compute agnostic. The founders and the team are all staying to keep pushing this mission forward.
Together, we think we can make open source the default way to build AI, with the goal of empowering 100 million AI builders to own their intelligence rather than rent it.
Excited about the next 10 years! 🤗🤗🤗
Aryan V S retweeted
Today, we are introducing Inkling.
Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available.
thinkingmachines.ai/news/int…
Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
Aryan V S retweeted
(1/6) The FastVideo team is excited to release a new set of FP4 attention kernels for B300, achieving up to 1.69x speedup over FA4!
Quick weekend post on how to glance inside your kernels and intra-kernel profiling with RMSNorm example 🧐
All work was done on the amazing @verdacloud thanks to generous compute for carrying out some exciting research. Lots more soon!
Link: a-r-r-o-w.github.io/blog/3_b…



















