What a Hearing Record Contains
Oral testimony, written statements, Q&A transcripts, and submissions for the record — what each component is, when it appears, and how it enters the SafeASI log.
Explainers & methodology
Explainers describe how the pieces fit together — what a compute threshold does, how a bill is structured, what a hearing record contains. They make no recommendations and take no positions.
Oral testimony, written statements, Q&A transcripts, and submissions for the record — what each component is, when it appears, and how it enters the SafeASI log.
What preemption clauses do mechanically — express vs. implied preemption, savings clauses, and how recent executive and congressional proposals treat state AI authority.
What new committee gavels actually change — agenda control, the oversight record, and veto points — and what stays the same. A procedural explainer, without predictions.
Why a FLOP count is the most consequential number in AI policy: how compute tripwires work, their four known failure modes, and how bills differ in addressing them.
Six questions that expose any AI bill's structure — the same dimensions the SafeASI tracker logs for every entry. Neutral methodology, no recommendations.
New explainers and every tracker update, in your reader of choice.