For what exactly? What risk are you referring to here?
One more party increase the risk of something going wrong. That’s just basic logic.
btw, Anthropic’s code was also just released (unintentionally).
Well, yes but also not really. The source code for Claude Code (their coding harness) got leaked, and it is quite informative for those of us who develop with or around AI.
What hasn’t been released is their backend/API code. We don’t know how the wiring on their server-side looks like and that’s where all the magic, telemetry, and “tracking for training” happens.
ah got it thank you.