Indeed a great article, as is the article from François he refers to.
What surprised me though, is that there are some very high scoring models (or better: runs) that didn’t reach the news, or at least the outlets I’m following.
Maybe I should get some AI-assistant to track everything 😉