reading

An MIT paper argues that the best way to regulate potentially risky technologies like AI is through gradual adoption while monitoring for negative impacts. via Marginal Revolution.

A chess experiment finds a human can beat Stockfish 14 with a queen advantage but not with just a rook. via LessWrong

Tyler Cowen shares a list of articles and videos, touching on AI, politics, natural resources, and music. via Marginal REVOLUTION.

Jacob Steinhardt explores what capabilities large language models could possess by 2030, such as superhuman coding, faster processing speeds, and training on diverse data beyond text and images. via LessWrong

Tyler Cowen shares a defense from Jared Diamond along with four other links on topics such as global inequality and Singapore's take on LLMs. via Marginal REVOLUTION.

Cory Doctorow's daily links cover a range of tech topics including content moderation, AI risks, privacy, and a security flaw impacting MSI computers. via Pluralistic: Daily links from Cory Doctorow