Dev says Google models are making real progress on lossless compression as intelligence proxy

gnukeith · x · 2026-11-22

The author reports significant progress on the "compression ≈ intelligence" problem using Google Neon and Astra 6.2, pushing lossless compression further — something he doubts he could do with open-weight models. He also runs a machine where agents operate 24/7 autonomously, saying Hy4 Flash has drastically improved his quality of life, and is waiting to try Qwen 4 Flash.

Original post →

More from Models

Models channel →