I expect this sort of thing fails to scale naively, but I'm also reasonably confident that a combination of 'wish indexing' with well-integrated networking agents can get *much* further in principle. We'll likely publish something on this soon as part of the 'design sketches for a more sensible world' (https://newsletter.forethought.org/p/design-sketches-for-a-more-sensible).
I'm also interested to see where this sort of thing could go for the n>2 case, which might get quite potent, a kind of evolution of change.org or kickstarter.
Thanks for your comment! So many exciting directions ahead
I agree that we can decouple the opportunity mining from opportunity filtering. The negotiation framing compresses two steps into one. It might still be relevant in cases where agents have access to private information that they can sometimes reveal when this could lead to better opportunities.
I'm also excited about the application of n>2 for hidden coalition discovery. I'm thinking of people sharing private information (e.g., political beliefs) with TEEs running coalition discovery agents. The only action coming out of the TEE is creating a "group chat" between people, saying "you have something in common in domain X." It's up to the members to decide how much they want to share.
I believe this could dramatically reduce situations where many people share an opinion, but nobody talks about it for fear of talking to someone who doesn't share it.
This is dope and love the "proposals". I could imagine that this would be cool for companies open to external partnerships.
Amazing! I've been background thinking about similar things for some months (e.g. you can see my comments on Nora's post: https://www.lesswrong.com/posts/mtASw9zpnKz4noLFA/gradual-paths-to-collective-flourishing?commentId=ytoiDjkp7eLNjxzKH). Great to see someone prototyping!
I think the real magic here is less about 'negotiation' and more about discovery and networking (or 'mining', as you put it) - something like an aligned recommender system specialised to finding opportunities for coordination and collaboration. There's a step 0: finding potential counterparties! https://www.lesswrong.com/posts/hqdvZGhxC45g57a6s/oliver-sourbut-s-shortform?commentId=AnhGvQvtmkYhMgSMj
I expect this sort of thing fails to scale naively, but I'm also reasonably confident that a combination of 'wish indexing' with well-integrated networking agents can get *much* further in principle. We'll likely publish something on this soon as part of the 'design sketches for a more sensible world' (https://newsletter.forethought.org/p/design-sketches-for-a-more-sensible).
I'm also interested to see where this sort of thing could go for the n>2 case, which might get quite potent, a kind of evolution of change.org or kickstarter.
Thanks for your comment! So many exciting directions ahead
I agree that we can decouple the opportunity mining from opportunity filtering. The negotiation framing compresses two steps into one. It might still be relevant in cases where agents have access to private information that they can sometimes reveal when this could lead to better opportunities.
I'm also excited about the application of n>2 for hidden coalition discovery. I'm thinking of people sharing private information (e.g., political beliefs) with TEEs running coalition discovery agents. The only action coming out of the TEE is creating a "group chat" between people, saying "you have something in common in domain X." It's up to the members to decide how much they want to share.
I believe this could dramatically reduce situations where many people share an opinion, but nobody talks about it for fear of talking to someone who doesn't share it.