Ex-Apple/Berkeley Researcher Launches Perceptual Labs for Weight-Space Programming
jzlegion · x · 2026-11-18
KJHMiao announces Perceptual Labs, betting that weight-space programming will shift AI from training models to engineering them — and eventually to self-improving models. Pitch: precise edits (change one behavior without forgetting), efficiency (edits written in language, no new training runs), and self-improvement (auditable weight edits). Use cases include model debugging via neuron tracing, behavior removal directly in weights, and capability transplant without new datasets. Waitlist is open.
More from Companies & People
- NVIDIA reportedly acquires Inception; diffusion model weights may go open source — mathemagic1an · 2026-11-18
- Chrome team hosts "Chrome Unprompted" builder meetup in London on Nov 24 — laparisa · 2026-11-18
- Vivek Haldar: hundreds of tiny enterprise tools are now cheaper to vibe-code than to maintain — vivekhaldar · 2026-11-18
- Noam Brown, father of superhuman poker bots, wins AI charity poker cup in SF — polynoamial · 2026-11-18
- Cohere and Korea's Upstage sign strategic partnership for sovereign AI — aidangomez · 2026-11-18
- The open-source AI landscape: NVIDIA+HF vs Anthropic's restrictions vs closed labs — tekbog · 2026-11-18