I think Scott Alexander made some good arguments in this recent post

"AI Safety Is Not A Corporate Plot"

https://www.astralcodexten.com/p/ai-safety-is-not-a-corporat...