Table of Contents

AI safety’s most underrated fic: Seed.​Starts off with a discussion about why they’re (trying to) shut down a supertintelligent AI.“Why? What are you afraid of? Viruses?”“No. We’re afraid of accidentally creating a merciless god.”“You guys can really do that? Why would you even do that?”  ​It also starts off by talking about interpretability and how we currently don’t understand AI and once it becomes superintelligent, it becomes super-unintelligible

It’s clearly written by somebody who understands ML and cares about AI safety and is really well written. Can find the whole webcomic here

Categorized in:

AI Risk & Governance,