Researcher argues AI shouldn't implement chip designs but architect them, with formal verification
satnam6502 · x · 2026-11-09
Researcher satnam6502, speaking at the IET REACH 2026 computer architecture conference in London, argues we shouldn't just use AI to implement architectures we already have in mind — instead specify behavioral requirements, let the 'alien brain' design freely, and formally verify output against the spec. He goes further: on FPGA-like platforms, conventional computer architecture itself may be dissolvable via partial evaluation.
Related event: IET Panel Debates Letting AI Design Chip Architectures From Scratch(2 posts)→
More from AGI Musings
- RL compute scaling: 10x more RL compute may double task horizons, exponent ~0.3 — tobyordoxford · 2026-11-09
- The real problem with conferences is unqualified reviewers, not LLM use — LucaAmb · 2026-11-09
- Toby Ord: RL scaling likely power law in capabilities, citing horizon-compute paper — tobyordoxford · 2026-11-09
- antirez: devs refusing AI coding help are visibly falling behind — antirez · 2026-11-09
- AI biweekly recap: METR finds 2x faster R&D at Anthropic, Apple cuts engineers' AI spend caps — emmanuelvivier · 2026-11-09
- Nobody Was Argued Into Loving AI: One Real Problem Converts Skeptics Better Than Any Debate — Pascallisch · 2026-11-09