Policy feasibility shouldn't dictate your view on AI risks
AndyMasley · x · 2026-08-31
The author critiques a cognitive pattern in AI safety debates where people judge the validity of a risk based on its policy implications. If an idea leads to ridiculous or unfeasible policies, people tend to dismiss the underlying risk itself. The author argues that risk assessment should remain independent of the difficulty of formulating a response.
More from AGI Musings
- Blind Refusal eval reveals models over-comply with authority directives — sethlazar · 2026-08-31
- YC President Garry Tan: AI Turns Markdown Files Into Employees, Enabling Tiny Teams — illscience · 2026-08-31
- Musk says AI and robots will solve the debt problem — McDonaghMatthew · 2026-08-31
- Critique: Alignment as a Capital-Friendly Narrative — viksit · 2026-08-31
- NYC Seminar: The Fragility of Borrowed Intelligence & AI Dependence — round · 2026-08-31
- Are we really living through a technology backlash? Demand decides tech success — technollama · 2026-08-31