A think tank to develop proof of stake for international conflicts
Artificial Intelligence, Great Power Relations, Space Governance
International conflicts pose a risk already, and that’ll only get worse when AI arms races start among countries. Yet, establishing a central world government is hard and bears the risk that it may be taken over by a dictator.
Currently we’re implementing an algorithm that puts at stake the lives of millions of citizens, and where almost anyone can slash the stake of almost anyone else. Instead we could put a lot of weath at stake and establish mechanisms for when the wealth of a nation gets slashed.
Various social services may be a starting point. Health care insurance, welfare, justice, and others don’t work very well in many countries. Various market mechanisms could be made available to people to use instead. People could pay regularly into an insurance pot that is at the same time lent out or invested to generate passive income for the firm and the insured person, and used to compensate anyone who the insured person might harm. (This is inspired by Hanson’s proposal for a tort law reform.) The insurers can collaborate with firms that aggregate judgements. These firms request judgements on cases form random judges and are themselves insured against biasing their random selection.
For big decisions, such as whether to slash the stake of a large group of people, such as a nation, many judges are needed. That also has the advantage that the judges can be far apart, to the point that it may take years for their judgments to travel at lightspeed to reach the nation they’re judging. Meanwhile the state of the nation can fork, which causes linearly more overhead in the number of forks as people need to do their transactions multiple times. But gradually more and more judgments will come in and will potentially resolve the uncertainty even before all judgments are in. That way it would scale better than a central government for all worlds.
I’m unfortunately pessimistic that this can be established in time to prevent AI arms races. But maybe it’ll turn out that we have more than a few decades left after all.
A think tank to develop proof of stake for international conflicts
Artificial Intelligence, Great Power Relations, Space Governance
International conflicts pose a risk already, and that’ll only get worse when AI arms races start among countries. Yet, establishing a central world government is hard and bears the risk that it may be taken over by a dictator.
Currently we’re implementing an algorithm that puts at stake the lives of millions of citizens, and where almost anyone can slash the stake of almost anyone else. Instead we could put a lot of weath at stake and establish mechanisms for when the wealth of a nation gets slashed.
Various social services may be a starting point. Health care insurance, welfare, justice, and others don’t work very well in many countries. Various market mechanisms could be made available to people to use instead. People could pay regularly into an insurance pot that is at the same time lent out or invested to generate passive income for the firm and the insured person, and used to compensate anyone who the insured person might harm. (This is inspired by Hanson’s proposal for a tort law reform.) The insurers can collaborate with firms that aggregate judgements. These firms request judgements on cases form random judges and are themselves insured against biasing their random selection.
For big decisions, such as whether to slash the stake of a large group of people, such as a nation, many judges are needed. That also has the advantage that the judges can be far apart, to the point that it may take years for their judgments to travel at lightspeed to reach the nation they’re judging. Meanwhile the state of the nation can fork, which causes linearly more overhead in the number of forks as people need to do their transactions multiple times. But gradually more and more judgments will come in and will potentially resolve the uncertainty even before all judgments are in. That way it would scale better than a central government for all worlds.
I’m unfortunately pessimistic that this can be established in time to prevent AI arms races. But maybe it’ll turn out that we have more than a few decades left after all.