It’s really tiring hearing the, “open source is bad for security!” argument whenever it pops up, but especially in domains that *have a lot of work to be done* when it comes to adopting open source, like LLMs and AI models.

Being MORE open is not the problem here. Opening up the weights and releasing under open licenses does not suddenly create security vulnerabilities, unless you believe solely in security through obscurity—which, I’m sorry, the tech industry has moved on from for the better.