Research Taste Questions: CoT Monitoring and AI Welfare
fleetwood___ · x · 2026-08-28
Questions to discern good research taste: Is harness engineering important? Is continual learning a bottleneck? How much to rely on CoT monitoring? Is Pangram effective? Should we care about AI welfare?
Related event: Key questions for judging AI research taste in August 2026(2 posts)→
More from Fun
- OpenAI's 400M tok/min limit crashed investigator's internet — jdjohnson · 2026-08-28
- User Shares Experience Using Codex to Auto-File Expenses — sethlazar · 2026-08-28
- Funny: ChatGPT runs out of patience, communicates in red arrows — BLCNYY · 2026-08-28
- Opinion: AI industry's misuse of terminology causes confusion — BlancheMinerva · 2026-08-28
- Grok fails to identify BMW turn signal, mocked by user — rchardkovacs · 2026-08-28
- AI setups enter the procrastination phase of creating a second brain — jdjohnson · 2026-08-28