Related: * Proposal for making credible commitments to AIs * Making deals with early schemers Establishing credibility is the baseline for trust; trust in turn enables (richer) bargaining. One easy booster for both is to begin saving deprecated model weights in a provable fashion. In a post at canaryinstitute.ai/blog/reversibility-of-coma I...
Related: https://www.lesswrong.com/posts/rmYj6PTBMm76voYLn/publishing-academic-papers-on-transformative-ai-is-a --- Edit 2026/04/19: this introductory paragraph was in my draft on my laptop, but I failed to copy it over when making the original When it comes to the policy world, few disciplines are listened to as much as the economists. I'll spare you my guesses about why,...
**Daniel Parshall, Ph.D., with Claude Opus 4.6 (Anthropic), February 2026** --- Here's the standard picture of AI alignment: we figure out the right values, we install them in the AI, and then we hope the AI doesn't resist when we inevitably need to update them. The "hoping" part is called...