Surveil Things Not People by Paul Christiano – Technology may advance to the point where individual people can destroy the world. In this scenario we may need pervasive surveillance. However we should surveil things, not people.
Honest Organizations by Paul Christiano – Goal: Create an honest AI organization. Strategy: The company makes public statements that it promises are not misleading. Employees commit to publicly dispute the statements if they find them misleading. This strategy cannot create trust but it can dramatically amplify levels of trust.
Techniques For Optimizing Worst Case Performance by Paul Christiano – “If powerful ML systems fail catastrophically, they may be able to quickly cause irreversible damage. To be safe, it’s not enough to have an average-case performance guarantee on the training distribution — we need to ensure that even if our systems fail on new distributions or with small probability, they will never fail too badly. The difficulty of optimizing worst-case performance is one of the most likely reasons that I think prosaic AI alignment might turn out to be impossible (if combined with an unlucky empirical situation). In this post I want to explain my view of the problem and enumerate some possible angles of attack. My goal is to communicate why I have hope that worst-case guarantees are achievable.”
Reid Hofmann Peter Thiel: Technology and Politics by Marginal Revolution – Link Post: “Cardinal Conversations: Reid Hoffman and Peter Thiel on ‘Technology and Politics’ moderated by Niall Ferguson”. Skip the first 15 minutes of intro.