PSI-0.6 released: one model reasoning in physical-world streams of thought
dyamins · x · 2026-10-27
The PSI team (Aw Khai Loong, Klemen Kotar et al.) releases Ψ-0.6, exploring the physical-world analogue of chain-of-thought: a single model that reasons in "streams of thought" across words, depth, motion, camera pose, and imagined scenes — and learns when to use which modality.
More from Multimodal
- Dan Shipper: video creation and game making will become foundations of knowledge work — danshipper · 2026-10-28
- Black Forest Labs lands $300M Series B as CEO pitches open visual models for robotics — bfl_ai · 2026-10-28
- BFL hosts 400-person Berlin party, shows FLUX-mimic driving a real robot at GTC — bfl_ai · 2026-10-28
- FLUX-mimic drives a real robot at NVIDIA's GTC Berlin booth — bfl_ai · 2026-10-28
- Turning gif artefacting into a stylistic edge for prettier AI art — zsakib_ · 2026-10-28
- Google's audio model IDs a song in a noisy café where Shazam fails — keenanisalive · 2026-10-27