anon 0x434 said in #2513 1mo ago:
Everyone seems to be trying to arms race their way to superintelligence these days. I have a different idea: supercoordination.
Supercoordination is when you use high powered (probably software-enabled) social technology to achieve unprecedented levels of social trust, agreement capability, intellectual good faith, trustworthiness of information, social information clearing, incentive alignment, etc. A supercoordinated group of people will be able to achieve their aims even against substantial resistance, and generally operate at a much higher level of social efficiency.
So here's the question: is there such a thing as software-mediated supercoordination, and what would that look like?
First of all, social technology is a lot more like a living organism than other tech. It grows, flourishes, gets cancer, and dies. This is the cultural cycle. It's not just rational and neutral tools, but systems of ends, pre-rational beliefs, taboos, foundational concepts.
For this reason unlike Yudkowsky (allah forgive me for uttering that name here) I don't believe supercoordination can be perfect. Even super advanced hyper-rationalist AI systems will wax and wane in their social and even self-coordination, and will never achieve any kind of Singleton. (And therefore there may still be individuated "people" who are in a state of semi-coordination and individual agency, even if not humans as we would recognize them).
The implication is that a new social technology stack that could achieve supercoordination has to be a new pre-rational culture and a political revolution, or at least ride such a wave. The advances in contract law, literacy, administration, professionalism, etc that led to the modern world were at first religious and political. Only later did they become cold and "neutral" techne. So we should not expect to be able to fully approach the question of social technology or supercoordination rationally and coldly. However, I still think it's useful to think through.
One imagination I have is a software protocol within some future superpolis that its citizens use to share information about themselves and things they want, do, and witness. This would be their legal system, political system, public forum, gossip network, etc. The protocol is designed so as to reliably route truth and reward around according to some scheme such that everyone gets their due reward for their deeds good and bad, everyone is incentivised to be a good citizen, and everyone has access to the common knowledge of each other's record of trustworthiness. Careful control of the flow of information, in some cases siloing things and in other cases maximally clearing the available information, would create the shape of their society and the particular scheme of trustworthiness and coordination they are operating on.
Depending how you imagine it, this could sound like some kind of panopticon horrorshow, or a high trust utopia. I prefer to think of it as the latter and try to figure out in more detail how to get there, though there are definitely implementations of this kind of thing that I would find horrifying and try to escape from. This is where the pre-rational leap-of-faith about what kind of life we want to live together is very important.
In any case I'm very curious to hear if anyone else has thought about this kind of thing, and has ideas of its potential and limitations, what it should look like, etc.
Supercoordination is when you use high powered (probably software-enabled) social technology to achieve unprecedented levels of social trust, agreement capability, intellectual good faith, trustworthiness of information, social information clearing, incentive alignment, etc. A supercoordinated group of people will be able to achieve their aims even against substantial resistance, and generally operate at a much higher level of social efficiency.
So here's the question: is there such a thing as software-mediated supercoordination, and what would that look like?
First of all, social technology is a lot more like a living organism than other tech. It grows, flourishes, gets cancer, and dies. This is the cultural cycle. It's not just rational and neutral tools, but systems of ends, pre-rational beliefs, taboos, foundational concepts.
For this reason unlike Yudkowsky (allah forgive me for uttering that name here) I don't believe supercoordination can be perfect. Even super advanced hyper-rationalist AI systems will wax and wane in their social and even self-coordination, and will never achieve any kind of Singleton. (And therefore there may still be individuated "people" who are in a state of semi-coordination and individual agency, even if not humans as we would recognize them).
The implication is that a new social technology stack that could achieve supercoordination has to be a new pre-rational culture and a political revolution, or at least ride such a wave. The advances in contract law, literacy, administration, professionalism, etc that led to the modern world were at first religious and political. Only later did they become cold and "neutral" techne. So we should not expect to be able to fully approach the question of social technology or supercoordination rationally and coldly. However, I still think it's useful to think through.
One imagination I have is a software protocol within some future superpolis that its citizens use to share information about themselves and things they want, do, and witness. This would be their legal system, political system, public forum, gossip network, etc. The protocol is designed so as to reliably route truth and reward around according to some scheme such that everyone gets their due reward for their deeds good and bad, everyone is incentivised to be a good citizen, and everyone has access to the common knowledge of each other's record of trustworthiness. Careful control of the flow of information, in some cases siloing things and in other cases maximally clearing the available information, would create the shape of their society and the particular scheme of trustworthiness and coordination they are operating on.
Depending how you imagine it, this could sound like some kind of panopticon horrorshow, or a high trust utopia. I prefer to think of it as the latter and try to figure out in more detail how to get there, though there are definitely implementations of this kind of thing that I would find horrifying and try to escape from. This is where the pre-rational leap-of-faith about what kind of life we want to live together is very important.
In any case I'm very curious to hear if anyone else has thought about this kind of thing, and has ideas of its potential and limitations, what it should look like, etc.
referenced by: >>2514 >>2583 >>2596 >>2597 >>2613 >>2680
Everyone seems to be
