I totally agree! I think this technology is likely to be the foundation of many future capabilities as well as safety. What I meant was that society is unlikely to replace today’s insecure and unreliable power grid controllers, train network controllers, satellite networks, phone system, voting machines, etc. until some big event forces that. And that if the community produces comprehensive provable safety design principles, those are more likely to get implemented at that point.
My point was more that I expect there to be more value in producing provable safety design demos and provable safety design tutorials than in provable safety design principles, because I think the issue is “people don’t know how, procedurally, to implement provable safety in systems they build or maintain” than it is “people don’t know how to think about provable safety but if their philosophical confusion was resolved they wouldn’t have too many further implementation difficulties”.
So having any examples at all would be super useful, and if you’re trying to encourage “any examples at all” one way of encouraging that is to go “look, you can make billions of dollars if you can build this specific example”.
I totally agree! I think this technology is likely to be the foundation of many future capabilities as well as safety. What I meant was that society is unlikely to replace today’s insecure and unreliable power grid controllers, train network controllers, satellite networks, phone system, voting machines, etc. until some big event forces that. And that if the community produces comprehensive provable safety design principles, those are more likely to get implemented at that point.
My point was more that I expect there to be more value in producing provable safety design demos and provable safety design tutorials than in provable safety design principles, because I think the issue is “people don’t know how, procedurally, to implement provable safety in systems they build or maintain” than it is “people don’t know how to think about provable safety but if their philosophical confusion was resolved they wouldn’t have too many further implementation difficulties”.
So having any examples at all would be super useful, and if you’re trying to encourage “any examples at all” one way of encouraging that is to go “look, you can make billions of dollars if you can build this specific example”.