While I generally lean toward the AI skeptical side, at least for more extreme claims on near-term LLM capability, growth and time frames, I'm not at all a fan of this. It seems like political grandstanding and unlikely to net much in the way of meaningful harm reduction.

If it goes anywhere at all, it'll likely just result in a settlement paid to the government and a consent decree mandating well-intended, nice-sounding yet vague rules which just become another compliance cost for leaders, barrier for emerging competitors and otherwise accomplish little of value for citizens. It's also unproductive because it tends to polarize a complex, nuanced and evolving technical issue toward extremes by hijacking it as fodder for existing political and even culture war battles.

Agree, I would much rather see meaningful and powerful regulatory action instead of silly lawsuits.