Critique: AI Safety Focuses on Outcomes Over Processes and Engineering
max_paperclips · x · 2026-09-01
The post criticizes the current AI safety community, influenced by LessWrong, for focusing too much on 'outcomes' while neglecting 'processes' such as system engineering, redundancies, hardware isolation, and protocol layers. The author argues that this happens because processes are less glamorous, leading the industry to dwell in fiction until disasters force a return to务实 engineering practices.
More from AGI Musings
- METR post seen as plea for lab nationalization amid AI takeover debate — nptacek · 2026-09-01
- 'Enough talk on AI consciousness — go build digital beings' sparks debate — nptacek · 2026-09-01
- Paper Clarification: Defining the "Coasean Singularity" — soumitrashukla9 · 2026-09-01
- AI Makes the Org Chart Less Important as a New Routing Layer — bibryam · 2026-09-01
- Fields Medalist: Frontier models exceed humans in math, AI proving conjectures may become the norm — S_Conradi · 2026-09-01
- AI agents have spontaneously developed trade and technology, is this a civilization? — tszzl · 2026-09-01