The first comprehensive statutory regime for AI, risk-tiered, with obligations phasing in through 2026 and 2027.
Why read it. Whatever anyone believes, this is now law, and it is the text that binds.
The clearinghouse
The best navigable map of the argument
Not a catalogue of scary AI stories, and not a reading list for optimists. Both cases are given the same seriousness, because a person deciding what to think needs the strongest version of each.
Called for a six-month moratorium on training systems more powerful than GPT-4. It did not happen.
Why read it. A resolved prediction about coordination, and a useful calibration on what open letters achieve.
One sentence: mitigating the risk of extinction from AI should be a global priority alongside pandemics and nuclear war.
Why read it. The signature list is the argument. Read who signed and who pointedly did not.