Come to UW and we’ll teach it to you. Our alum Matt Deitke (4 yrs BS, dropped out of PhD after 1 yr, few months at startup) went straight to making $250M at Meta.
RESEARCH
-

GLM-4.5: Agentic Reasoning and Coding Foundation Models
By
–
GLM-4.5 Agentic, Reasoning, and Coding (ARC) Foundation Models
-

Voost: Unified Scalable Diffusion Transformer for Bidirectional Virtual Try-On
By
–
Voost A Unified and Scalable Diffusion Transformer for Bidirectional Virtual Try-On and Try-Off
-
Half of NSF budget wasted remains unknown mystery
By
–
Half of the NSF budget is wasted. We just don’t know which half.
-
Cutting NSF Funding Would Harm the United States
By
–
A major cut to the funding of the National Science Foundation would be very bad for the future of the US. https://thehill.com/opinion/technology/5439205-congress-wants-to-cut-the-smartest-investment-taxpayers-ever-made/ [Translated from EN to English]
→ View original post on X — @geoffreyhinton, 2025-08-11 02:45 UTC
-
OpenAI’s GPT-5 Release Timing Dilemma: Innovation versus Perfection
By
–
Sam Altman’s dilemma: OpenAI could have waited to release GPT-5 until it had something really worth bragging about, but that might have been forever.
-

InfiGUI-G1: Adaptive Exploration Policy for GUI Grounding
By
–
InfiGUI-G1 Advancing GUI Grounding with Adaptive Exploration Policy Optimization
-

Qwen3-4B First Model to Reject Unrealistic Test Requests
By
–
Qwen3-4B-Thinking (a 4GB download) just became the first model I've tested to directly push back against my pelican riding a bicycle test, calling it "oddly specific and completely unrealistic" and saying "this request violates physics and biology" https://
simonwillison.net/2025/Aug/10/qw
en3-4b/
… -

JAX Performance Meets Keras 3 Developer Velocity
By
–
JAX = performance & scalability Keras 3 = high velocity development, compact code, best practices by default Both at the same time = pretty killer