Alignment Plan
edit: see update:
everything below is somewhat outdated
Learn how to help individuals grow seamlessly, radically, rapidly.
Scale the process with AI (if necessary).
Unblock global coordination and alignment?
1. Help individuals grow
Become really, really good at causing clients to quickly untangle their biggest blocks (and without disrupting their life).
I’m focusing on people who are highly intelligent, AI/tech-adjacent, and eager to improve their emotional health.
Key assumption: Rapid, radically effective growth is possible.
2. Scale with AI
Scale my expertise by creating a voice-to-voice AI facilitator (if necessary).
Start with augmentation, end with automation: First, augment my service by giving clients access to an AI facilitator to use when I’m not available. Over time, improve the facilitator until most can use it without ever needing to talk to me. (The final product may be an AI- coach, game, social activity, or something else entirely.)
Current efforts: I am systematizing the process of growth. I’ve also made AI prototypes, but they’re not very good— They would be much better with end-to-end voice models like GPT-4o (see here).
Key assumption: There is no easier way to scale growth without AI. Edit: But it might easily be the case that a few human coaches is enough.
Caveats: Alex Zhu (mentor of mine; endorsed by Emmett Shear) emphasizes that I must ensure that the AI tool becomes a true prophet and not a false prophet. He also cautions that creating a perfect version of this tool is itself a “metaphysics-complete” problem. Alex has some related thoughts here.
3. Unblock global coordination?
Speculative.
Offer rapid, radically effective growth as a service at scale, probably facilitated by AI facilitators. Unblock global coordination / AI alignment by focusing on decision makers and engineers in AI.
Current effort: I already have clients who work in AI / alignment / policy.
Key assumption: If the individuals involved in alignment grew, that will be good for future alignment outcomes.
Some reasons I suspect this is true:
I noticed that many social conflicts would probably dissolve if the individuals involved grew. (I think many of the conflicts around global coordination and AI alignment are no different.)
Alignment is a metaphysics-complete problem and individual growth often unblock metaphysical insights. For example, growth can help one understand open individualism, Newcomb’s Problem, causal distance/boundaries, and self-fulfilling prophecies.
Thanks to Alex Zhu and Ivan Vendrov for mentorship. Thanks to the Center for Applied Rationality, Alex Zhu, Blake Borgeson, Nolan Kent, and anonymous for incidental funding while I developed these ideas. Thanks to Stag Lynn, Kaj Sotala, Brian Toomey, Anna Salamon, and Damon Sasi for collaboration.
Would you like to fund my work? Email me. Retroactive funding preferred.