Ezra Klein argues that policy debates on artificial intelligence fixate on risks—job loss, surveillance, bioweapons, wealth concentration—while neglecting how AI can serve the public. He notes that despite exploding usage, AI enjoys low public approval and even religious condemnation. Klein urges legislators to allocate funds for data centers and open‑source datasets, ensuring equitable access.

Examples illustrate AI’s upside. A OpenAI model solved an 80‑year‑old mathematics conjecture, while DeepMind’s Graphcast accelerated weather forecasts beyond supercomputer speed. The 2024 Nobel Chemistry prize honored AlphaFold, which relied on the publicly funded Protein Data Bank. Klein stresses that such breakthroughs required massive compute and curated data, resources unavailable to most universities without government support, without such backing, progress stalls.

Klein calls for a public AI option—a frontier‑level model owned by the state and funded to keep compute affordable for academia and public agencies. He suggests pairing this with guaranteed markets for outcomes, echoing Operation Warp Speed, to lure private firms into tackling rare‑disease drugs or new battery materials. A concrete public AI platform would turn the private‑public compute gap into an asset.