/explore
|
You are here 
|
morganredfield.com | ||
| | | | |
www.greaterwrong.com
|
|
| | | | | Eric DrexlerCentre for the Governance of AIUniversity of Oxford This document argues for "open agencies" - not opaque, unitary agents - as the appropriate model for applying future AI capabilities to consequential tasks that call for combining human guidance with delegation of planning and implementation to AI systems. This prospect reframes and can help to tame a wide range of classic AI safety challenges, leveraging alignment techniques in a relatively fault-tolerant context. | |
| | | | |
www.lesswrong.com
|
|
| | | | | Suppose you know what outcomes are better than what other outcomes, but not by how much. Given two possible actions you could take,AandB, if you k... | |
| | | | |
reasonableapproximation.net
|
|
| | | | | [AI summary] The author critiques and clarifies the Kelly Criterion by arguing it optimizes rank-based growth rates and expected log-money rather than acting as a general utility function, debating how this mathematical framework applies to philanthropy and charity. | |
| | | | |
www.alignmentforum.org
|
|
| | | Summary The core ideas that constitute the multi-disciplinary view[1]on AI safety research are: ... | ||
