The publicisation of the HF incident is currently leading to some amount of backlash in the general (not AI-focused) tech community, in public and private spaces.

For example, Bryan Cantrill wrote this post writing off all existential risk as fearmongering, and essentially equating its spread with "pranking people": https://bcantrill.dtrace.org/2026/09/13/the-contagion-of-fear/

Posts like this are hitting top positions on link aggregators, and are getting very mixed feedback, with one of the trending positions being that talking up the HF incident etc. is just pre-IPO marketing of the AI companies.

Further, I work at a large tech company that is not a frontier model developer, but we do have in-house pre-/post-training on large GPU fleets, and our "alignment" team is working on making sure that the models use tool calls frequently enough. This seems to be similar in other companies in the same spot (not top tier, but still resourceful): "alignment" is watered down to something like "instruction following" and that's it.

I'm worried that actual risk discussion becomes something on the level of fringe views like, I don't know, ancient civilisations or UFO conspiracies, as it becomes known in the mainstream but not well enough to be taken seriously.