So do frontier labs just sit on checkpoints and all release them strategically? Or is the advancement coincidentally just that fast?
1
91
Bears, beets, battle hymn of the republic
This will make you smile. Rain Wilson shares his observation of returned missionary reunions at the SLC airport.
Mystic Observes Book of Mormon Promise piped.video/shorts/bwzmjAPKo… via @YouTube
8
638
Should a self-driving car:
67%
Hit a tree, kill a driver
33%
Hit people, save driver
9 votes • Final results
1
170
My primary realization when I first saw TypeSafe's demo this week is that 90-95% of an agentic harness is just classification.
2
207
WATCH: @MightyHeaton hates the reasons you hate data centers.
30
163
942
155,624
Pretty exciting for Andrew and the Cartwheel team. Cartwheel just discovered that 3D human motion has a scaling law. This puts them in the same league as the frontier AI labs, but in a totally new domain.
Presenting compute-optimal scaling laws for human motion generation!!!
Human motion generation has historically been niche and difficult to scale. Animators, roboticists, motion analysts, and game developers all see the potential, but it rarely lives up to the rigor required in production.
We have long believed this is because all prior systems have been fantastically data constrained. Until today, we haven't been able to prove that scaling would actually help.
Today we present our flagship research paper: Compute-Optimal Scaling Laws for Human Motion Generation. (link: github.com/Cartwhl/scaling-l…)
We are building the world's largest, high-quality human-motion dataset. We trained hundreds of models across varying scales. They match the Chinchilla scaling found in language.
This places motion as the fifth data modality we understand how to scale. Text, images, audio, and video all have scaling laws. None of them really reflect the physical world. 3D human motion, across time, does.
Both autoregressive and flow matching motion models scale predictably. We can now predict, before training a model, how good it will be.
Motion is no longer data constrained. It is now a predictable and scalable frontier for physical and digital intelligence. Our largest models are training now, and the gap between generated motion and the real physical world is closing fast.
paper: github.com/Cartwhl/scaling-l…
blog: getcartwheel.com/blog/scalin…
2
6
559
Nate Sanders retweeted
🔥 Wildfire smoke plumes were easy to spot FROM SPACE today.
Utah now has five active wildfires burning, with four more just across the border in Arizona, Nevada, and Idaho. Watch how these fires rapidly grew throughout the day in this satellite loop. 🔥
35
135
1,062
125,487
Everything you need to know about the veracity of quantum computing is insinuated and illustrated in this post.
2
1
404
There are no Noams left. They have all the Noams.
I’m excited to share that I’ll be joining OpenAI and look forward to working with the exceptional team there.
It was a difficult decision to move on. I’m incredibly proud of the amazing team at Google and everything we’ve built together. It has been an honor and a pleasure to work with all of you.
3
279
I have been living righteously in hopes this moment would come.
The Legend of Zelda: Ocarina of Time will be reborn on Nintendo Switch 2 in 2026. #NintendoDirect
1
5
469
.@DarioAmodei I need you to ask your mechanistic interpretability team to get to the bottom of why Opus is always saying, “belt-and-suspenders.” Thank you for your attention to this matter.
1
3
212
Nate Sanders retweeted
non-designers who have never designed anything: "designers are cooked!"
people who have worked in code, want to work in code, always will work in code: "future of design is code!"
companies selling tokens: "move that button 4px with a prompt!"
designers:
86
141
1,543
72,468
