Matt Zhou

@Mattzh1314

tinkering

San Francisco, CA

Joined January 2013

332 Following

571 Followers

1 Subscription

Replies Reposts Media

Pinned
Matt Zhou
@Mattzh1314
May 15
We realized last year that we couldn’t train the models we wanted to train without the right type of data…so we made the data. Had a wonderful time with the folks from Cambridge/Oxford @RayLi234 @XiaoyangLyu22 and @elliottszwu! And shoutout @Remotion for helping make the video
RuiningLi
@RayLi234
May 15
🚀 Introducing Articraft, a coding agent for articulated 3D asset creation. Articraft writes code, executes it, receives validation feedback, and refines the result into simulation-ready 3D assets with parts, joints, and motion. We’re also releasing Articraft-10K: 10,000+
00:00
4
Matt Zhou
@Mattzh1314
Oct 16
BY @AndiGu00
Anthropic
@AnthropicAI
Oct 16
New from our STEM Fellows Program: Designing more efficient quantum circuits with Claude. Given plain-language instructions, Claude Mythos 5 designed quantum circuits that improve on the best published ones, sometimes by an order of magnitude or more. More efficient circuits
Matt Zhou
@Mattzh1314
Oct 12
LLMs 🤝 physical world
Chuanxia Zheng
@ChuanxiaZ
Oct 12
When I joined Articraft project, I thought agents could generate only native 3D assets. Led by Matt, we discovered that they could also create impressive articulated objects—but these assets were still difficult to use in robotics. This gap motivated ARTISAN. ARTISAN moves us
Quote
RuiningLi
@RayLi234
Oct 12
Agentic real2sim2real is going to unlock many long-horizon capabilities for robotics. In this work led by Xudong, we aim to make agent-generated assets more "physically correct", with less self-penetration and better performance in downstream robotic tasks! A couple of questions
Matt Zhou
@Mattzh1314
Oct 2
One thing that highlights the difference between Jev and an LLM w/ constrained decoding is that it is much better at assigning consistent scores when there are equally valid answers. [1/4]
1
Matt Zhou
@Mattzh1314
Sep 18
can someone send this to respawn so they can nerf aim assist
Flyte pablo.minecraft.2011
@pablominceraft
Sep 16
The Fruit Fly can play axle on controller I trained all 166,700 neurons in the MaleCNS v1.0 fruit fly connectome to play Axle in Apex Legends, using YOLOv5 plus video pretraining. With aim assist enabled, most of the neurons were unnecessary. #ApexLegends
00:00