The past several years have seen a remarkable improvement in fiscal responsibility within the effective altruism movement. The post Getting it right (Part 5: Fiscal responsibility) appeared first on Reflective altruism .
A second shutdown theorem due to Krakovna and Kramar is also unconvincing. The post Revisiting the shutdown problem (Part 4: Krakovna and Kramar) appeared first on Reflective altruism .
A leading shutdown theorem due to Elliott Thornley needs strong unmentioned assumptions to ground shutdown-resistance. The post Revisiting the shutdown problem (Part 3: Thornley) appeared first on Reflective altruism .
Informal arguments for Catastrophic Shutdown Difficulty include the Argument from Instrumental Convergence and the Empirical Argument. The post Revisiting the shutdown problem (Part 2: Informal arguments) appeared first on Reflective altruism .
The catastrophic shutdown problem is (roughly) the problem of designing AI systems that can be shut down when their acts would lead to existential catastrophe. The post Revisiting the shutdown problem (Part 1: Introduction) appeared first on Reflective altruism .
Anti-fanatical decision theories have problems of their own. The post Fanaticism (Part 4: Problems for anti-fanatical decision theories) appeared first on Reflective altruism .
There are many reasons to doubt the truth of fanaticism. The post Fanaticism (Part 2: Further reasons for doubt) appeared first on Reflective altruism .
The case for longtermism is strongest on fanatical decision theories. They might be wrong. The post Fanaticism (Part 1: Introduction) appeared first on Reflective altruism .
Yudkowsky and Soares have a simple solution to AI safety: Shut it down. The post If anyone builds it, everyone dies (Part 5: Solutions) appeared first on Reflective altruism .