Machine Learning is Linear Algebra

Short overview of me:

I am the CEO and co-founder of Apollo Research. I'm currently primarily thinking about AI model evaluations ("evals"). I'm excited about the benefits that AI can create but to ensure humanity benefits from AI, I primarily work on reducing the downside risks of AI systems. I'm particularly interested in the problem of scheming, where AI systems covertly pursue misaligned goals.

For more details, see my ‘CV’. Not all of my writing is on this blog. You can find most of my latest work on the Apollo website.

Evals

I want to build and grow the evals field. In the Evals section I list relevant evals resources and describe my field building plans.

Machine Learning Projects:

In the ML projects section I discuss selected ML projects.