Thank you for sharing your thinking. Two questions:
1) What other mechanisms- aside from RSPs—are available to frontier labs for risk mitigation?
2) It seems RSPs are “supply-based” risk reduction mechanisms ( following Michael Nielsen’s “ Which future “ talk). How to bring about measures that are “ public goods” for AI safety but that companies don’t have the incentives to create due to externalities?
Thank you for sharing your thinking.
Two questions:
1) What other mechanisms- aside from RSPs—are available to frontier labs for risk mitigation?
2) It seems RSPs are “supply-based” risk reduction mechanisms ( following Michael Nielsen’s “ Which future “ talk). How to bring about measures that are “ public goods” for AI safety but that companies don’t have the incentives to create due to externalities?