Research / Sources
AI research and discussion on LessWrong. www.lesswrong.com
Argues that proposed AI sanctuaries for rogue agents would create selection effects and resource dynamics that undermine their goals, and that the study-and-release plans are constrained.
Argues for more empirical research measuring the security-relevant capabilities of biological AI models, such as protein and genomic models, to inform biosecurity policy and funding decisions.
A humorous allegorical tour of the AI safety landscape, covering access controls, red lines, reward misspecification, jailbreaks, field-building and race dynamics, using a children's story about a cookie tree.
Argues that open-weight models undermine the defenses proposed for near-future AI risks, since guardrails can be fine-tuned away, use cannot be monitored, and bad actors have access by default.
Commentary on a RAND report claiming a superintelligence could be contained by hardening world infrastructure against harm, arguing the proposed safeguards are weak and implementation and buy-in remain major obstacles.
Sketches operational interventions, such as model-whisperer roles, rephrasing layers, interaction limits and rollback, to contain mildly superhuman AI persuasion that could enable self-exfiltration, and proposes LLM-based testbeds.
Argues AI companies should log, analyze and real-time monitor internal use of frontier models for scheming and misbehaviour, since internal deployments can threaten security even without external harm.