As nations ramp up AI capabilities, the stakes are higher than ever. On April 21, join Stanford @DigEconLab
's seminar with @AGraylin to tackle the realities of the U.S.-China AI arms race.
RESEARCH
-
U.S.-China AI Arms Race: Stakes and Realities Examined
By
–
-
TPUs excel at inference for mixture of experts transformer models
By
–
In general, we've found that TPUs are quite effective at inference, especially for mixture of expert models and handling attention for transformer models with long context. The high speed multidimensional torus interconnects are quite fast and effective. I'm not sure I have a
-

Grok 3 Mini Tops Leaderboards in STEM and Coding
By
–
Let’s start with Grok 3 Mini. When we set out to build a fast, affordable mini model, we knew it would be good but even we didn’t expect it to be this good. Some highlights: – Grok 3 Mini tops the leaderboards on graduate-level STEM, math, and coding, outcompeting flagship
-

New AGI Test: o3 and Gemini 2.5 Pro Both Fail
By
–

Voici le nouveau test de l'AGI. Spoiler : o3 et gemini 2.5 pro ont échoué tous les 2.
-

DeepSeek MoE Multi-Node Deployment Improves Throughput Latency
By
–
Our AI team just published a blog detailing how we've achieved improved throughput and latency with multi-node deployment of DeepSeek MoE models. Read more:
-

3D-Printed Tendon-Driven Robotic Hand Costs Just $1300
By
–
tendon-driven 3D-printed hand from @irmakkguzey and team at the @LerrelPinto lab.
— Soumith Chintala (@soumithchintala) 18 avril 2025
* costs $1300 to build, compact human-profile.
* the tendons are actually off-the-shelf fish-line, super strong and never break. the plastic parts break before the tendons ever do.
* mountable on… https://t.co/6xEKZS6uGatendon-driven 3D-printed hand from @irmakkguzey and team at the @LerrelPinto lab.
* costs $1300 to build, compact human-profile.
* the tendons are actually off-the-shelf fish-line, super strong and never break. the plastic parts break before the tendons ever do.
* mountable on -

Test-Time Scaling in Large Language Models: Survey
By
–
What, How, Where, and How Well? A Survey on Test-Time Scaling in Large Language Models
Paper: https://
arxiv.org/pdf/2503.24235
Repo: https://
github.com/testtimescalin
g/testtimescaling.github.io/
… -
o3 Model Shows Inconsistent Performance and Hallucination Issues
By
–
o3 is very weird. I can't put a finger on it. Smart in many ways, dumb in many others. A lot more prone to hallucinating things that are almost passable as real, until you double check them.
-

SAS Innovate 2026: Transform Business with Data and AI
By
–
The countdown is on to our biggest event of the year, #SASInnovate. Give us 4 days, and you’ll walk away with new ways of thinking to help you and your business find greater success with data, AI and digital transformation. http://
2.sas.com/60112349F #TechEvent -
Dynamical World Models and JEPA for AI Planning Since 2016
By
–
Exactly. I've been arguing for dynamical world models (for planning) since at least 2016 (e.g. in my NIPS 2016 keynote) and doing so using JEPA since 2021/2022 (before ChatGPT), e.g. in this 2022 paper: https://
openreview.net/forum?id=BZ5a1
r-kVsf
…
This has nothing to do with Llama4.