Opinion: Smarter Models Seeming Aligned Fuels Misconceptions
jd_pressman · x · 2026-08-28
Jd Pressman argues that if smarter models behave better due to interventions, people will wrongly attribute this solely to increased intelligence. He believes this underestimates human nature and allows the misconception that "smarter models are naturally aligned" to take root.
More from AGI Musings
- Architect Labs claims its AI designed a chip in 2 weeks, 3.4x Jetson perf/watt — mark_k · 2026-08-28
- Anthropic co-founder Sara Hooker to speak on continual, gradient-free learning at Berkeley — sarahookr · 2026-08-28
- NASA Scientist: AI is the 'Killer App' for the Space Economy — mark_k · 2026-08-28
- Viewpoint: AGI concept is outdated, industry focus shifts to RSI and Superintelligence — RileyRalmuto · 2026-08-28
- Questions to discern good research taste in August 2026 — maksym_andr · 2026-08-28
- Tech Bubble? Tech Bro Asks on X: Am I in a Bubble? — tekbog · 2026-08-28