Altman to Politico: Let's accept "some bad things" for the benefits of AI

Altman to Politico: Let's accept "some bad things" for the benefits of AI
Sam Altman told Politico Decoded (Oct. 4) that the world must accept "some bad things" for the benefits of AI, advocating light regulation and distancing himself from Anthropic. Reuters and the press pick it up today.
The full sentence: benefits justify accepting risk; technology must remain widely accessible; and it won't trade profits for "zero scams or zero hacks" — though it draws the line at catastrophic risk. Calculated transparency: It also promises to disclose more rogue incidents (none like the ones already public, it says).
You may also likeEl CEO de Microsoft dice que alquilaría capacidad de IA de Musk en el espacio
The contrast is Amodei: curbing and regulating with explicit government policing. Altman offers accessibility + voluntary disclosure. Two lab theologies in head-on collision, with OpenAI's IPO in the background (confidential docket, $60B+).
Critics read dated cynicism: calling for tolerance for harm the same week of safety layoffs and $500K/day audits. Advocates read weird honesty: no one else says out loud that progress has victims.
The nuance that matters: "accept" who? The risk is shared by users, hospitals and electricity networks; the benefit is capitalized by the captable. Without this asymmetry on the table, the debate is theater.
What follows: if the promise of transparency is fulfilled in real incidents and if Congress buys the "light touch" or legislates the catalog of "bad things" allowed.
You may also likeTrump amenaza con 'impuesto tremendo' a data centers de EE.UU. en el extranjero
Image credit: Sam Altman at TED, S. Jurvetson via Wikimedia Commons, licensed under CC BY 2.0.