Ten people on the inside — LessWrong

(Many of these ideas developed in conversation with Ryan Greenblatt) …

x Ten people on the inside — LessWrong AI Frontpage 2025 Top Fifty: 14 % 155 Ten people on the inside by Buck 28th Jan 2025 AI Alignment Forum 5 min read 28 155 Ω 64 (Many of these ideas developed in conversation with Ryan Greenblatt) In a shortform , I described some different levels of resources and buy-in for misalignment risk mitigations that might be present in AI labs: *The “safety case” regime.* Sometimes people talk about wanting to have approaches to safety such that if all AI developers followed these approaches, the overall level of risk posed by AI would be minimal. (These approach

saved by

• Elizabeth Qiu
• Anastasia Wei

related reading

• The case for ensuring that powerful AIs are controlled — LessWronglesswrong.com
• Ten people on the inside - by Buck Shlegerisredwoodresearch.substack.com
• Anthropic's leading researchers acted as moderate accelerationists — LessWronglesswrong.com
• Plans A, B, C, and D for misalignment risk — LessWronglesswrong.com
• The Case Against AI Control Research — LessWronglesswrong.com
• AI Control: Improving Safety Despite Intentional Subversion — LessWronglesswrong.com
• AI safety undervalues founders — LessWronglesswrong.com
• Current AIs seem pretty misaligned to me — LessWronglesswrong.com
• Ten AI safety projects I'd like people to work on — LessWronglesswrong.com
• My picture of the present in AI — LessWronglesswrong.com
• Ten AI safety projects I'd like people to work onthirdthing.ai
• The case for ensuring that powerful AIs are controlledblog.redwoodresearch.org