Prime Intellect's @xeophon argues open-source AI and safety aren't contradictory: "It is important that open source does not mean that safety is not a thing, which we have shown with that blog post and that we will continue to work on pretty heavily." "The safety work itself needs to be open. A lot of monitoring information at labs, how it works, findings, is pretty closed off. If you find some, it's highly summarized with little insight into what actually happened." "As more and more…