White House excludes open models from framework to test advanced AI capabilities

Thanks. The language in that article is a bit unclear/contradictory, but my read of it was actually the only the big 3 U.S. labs would be subject to it for now:

> only makers of closed, proprietary U.S. models that demonstrate state-of-the-art capabilities in cybersecurity and hacking based on performance benchmarks would have to voluntarily submit those models to the government for testing before they are released

(Separate from who is affected, even the language around “having to” “voluntarily” submit seems contradictory - if you “have to” submit then it isn’t ‘voluntary’.)

Axios has lost a lot of credibility as of late. Seems like they're quite willing to journalism wash anything the current dministration tells them.