"Think of the kids!" OK but the impact of this technology shouldn't be underestimated and one doesn't only have to read up on the extreme ends. From personal experience, have you ever "chatted" with a model outside of technical/work related stuff? Both Anthropic as well as OpenAI models have been overstepping personal (user) boundaries for years now. There are multiple reasons for why Google hasn't capitalized on the transformer much earlier..
I'm neither a proponent of the recent Zuckerberg maneuver to "protect the children!" nor of the EU based zero-knowledge proof track, but the people in those companies need to be kept accountable.
The new Claude availability constraints seem to me then like a (late) hedge against any major litigations on that front..
https://en.wikipedia.org/w/index.php?title=AI-induced_psycho...