This site and other resources

Since 2023, I've focused on writing about what AI companies are doing in terms of safety.

AI Lab Watch

My main project is AI Lab Watch. It's essentially a scorecard evaluating whether AI companies are doing what they should do in terms of safety. It's broader than this site; this site has much more detail on evals.

This site

On this site, I analyze five companies' evals: how good they are and whether they support the companies' claims. (I'm also experimenting with adding analysis of companies' plans and practices for safety and security.) Click a logo below to go to my page on that company.

Anthropic Logo
OpenAI Logo
Google DeepMind Logo
Meta Logo
xAI Logo

Hover over a logo to see a brief summary.

Click to go to the full page.

(There is also a little work related to this site by others.)

My main site content is primarily aimed at high-context professionals. My blog aims to be accessible to readers with basic context. I recommend casual readers subscribe to my blog for updates, then maybe try browsing the site in case they find it helpful but not try to read much of it. The blog will be about half summaries of analysis on companies' new reports and half other writing on what companies are doing.

Get updates

Get emails with analysis on companies' new reports