alt Hacker NewsTimnit seems to mostly think it's dangerous because of things other than the model itself, e.g. AI labs' political influence and data centre resource consumption (mentioned in the article). The core thesis seems to be "wake up and stop wasting so much resources on this useless parrot".
She isn’t as big on P(DOOM), but she does worry about things like biological weapons or autonomous war vehicles. Her take is that they are worse than useless, but can be actively harmful.