One of the most exciting findings in our @Nature paper is the discovery of a clear scaling law of AI science. By using our Automated Reviewer to grade papers generated by different foundation models, we observed that as the underlying models improve, the quality of the generated

→ View original post on X — @sakanaailabs