Terence Tao: Quantitative AI progress needs accurate and transparent evaluation

To use the Mastodon web application, please enable JavaScript. Alternatively, try one of the native apps for Mastodon for your platform.