GPT-6 Astra is more aligned than our previous models. But it’s also less monitorable, which is a concerning trend that we take very seriously. We believe monitorability drop comes from a jump in intelligence and not direct optimization pressure on CoT or architecture changes. More thoughts in the thread.
🚨 “The Diffusion Duality” is out! @ICML2025
⚡️ Few-step generation in discrete diffusion language models by exploiting the underlying Gaussian diffusion.
🦾Beats AR on 3/7 zero-shot likelihood benchmarks.
📄 Paper: https://t.co/vyl1mVVI81
💻 Code: https://t.co/LN883NB8nr
🧠 Blog: https://t.co/3XMb2xfxLz
(1/8)
Two cool papers from yesterday:
- MMaDA: Multimodal Large Diffusion Language Models https://t.co/nh1loZZLs3
- Web-Shepherd: Advancing PRMs for Reinforcing Web Agents https://t.co/DPzFYmidOM
"I've heard from several people that they've consolidated their companies tool dependencies into a single package and this worked really well for them." - @dan_abramov
My team did this year ago. 100+ packages, build, lint, test config in 1 package!👍