A specimen catalogue of the highest-leverage work

Every idea a person might give a life to.

After Halcyon Molecular, Michael Andregg ran a deliberately wide search for the single highest-impact thing to work on — a running list of speculative projects across AI, uploading, cryonics, nanotech and biology, each scored on impact, tractability and neglectedness. That search led to AI compute (and to Fathom) as the best lever for AI safety. This is the list.

Draft · v0 · merged from the 2017–2021 lists · not all entries are endorsements — it’s a map of the possibility space.
0projects