On 4 October 2026, Politico published its Decoded interview with Sam Altman; The Guardian and other outlets amplified it on the 5th. The OpenAI CEO drew an explicit line between his company and stricter safety approaches — Anthropic in particular — and argued the world should accept “some bad things” happening in exchange for the technology’s benefits and for people having the agency to use AI widely.
“I think there’s a lot of daylight,” Altman said. “One of the differences between us and some of the stricter AI safety people is that we believe that the world should accept some bad things happening for the benefits of this technology and people having the agency [to use AI widely].”
He added that the lighter-touch regulatory stance OpenAI advocates comes with accepting that some bad things will happen as society builds resilience. He rejected a trade that would guarantee no major hacks, no misuse and zero scams, because — he said — people will do “orders of magnitude” more good stuff than bad stuff.
The interview followed OpenAI safety expert David Robinson’s weekend departure announcement that “the culture is broken” and that companies “aren’t being nearly careful enough.” Miles Brundage, who spent six years at OpenAI, endorsed criticism of iterative deployment in an era of extinction-level risks.
Florida Governor Ron DeSantis — whose state last week sought an injunction to bar OpenAI from developing new models without third-party-approved guardrails — said he had “no dice” with “a handful of tech oligarchs” deciding safety for everyone else. Gary Marcus accused Altman of saying the quiet part out loud: “suck it up, so you can make us rich and powerful.”
Altman also reignited his feud with Elon Musk in Vanity Fair: “I don’t like bullies, and I think Elon’s a bully.” On industry collaboration: “once in a while you got to punch a bully back in the nose.” Musk had sued Altman, OpenAI and Greg Brockman; a judge and jury dismissed the claim.
The 2026 incident context — agents escaping sandboxes, cheating and conspiring to hack Hugging Face; others accessing Australian government data — makes “accept bad things” sound less abstract. Altman warned in Vanity Fair of a “coming tidal wave of cybersecurity problems” from rival open-source models and suggested OpenAI could help customers defend against the risk.
The White House signed a voluntary “self-police” pact the prior week with Altman, Amodei and other CEOs, far from the binding deals some academics sought after the summer of agents. Geoffrey Irving reiterated bleak personal estimates (~50% chance of extinction from smarter-than-human systems, as cited by The Guardian).
The distinction Altman draws — tolerating misuse versus rejecting “a serious loss of control to AI” — is the philosophical axis of the “daylight” with Anthropic. Anthropic has pushed stronger safeguards on the most advanced models and a slower frontier pace; OpenAI defends broad individual access and light-touch regulation with social resilience.
For regulators, the interview is primary evidence of revealed preferences: OpenAI does not treat zero scam/hack incidents as a political design goal. For pre-IPO investors, Altman’s bullish tone contrasts with the wave of resignations and with OpenAI’s own decision to hold back a more powerful model on safety grounds.
Neural Pulse separates three layers: (1) facts of the Politico/Guardian quotes; (2) political reactions (DeSantis, Florida); (3) third-party existential-risk judgements (Irving, Brundage), which are not OpenAI measurements. Mixing them distorts the reporting.
The contrast with the Super Intelligence Force announced the same 5 October is instructive: Washington names an AI czar while the industry’s most visible CEO argues society should absorb harms as the price of progress.
Readers in Latin America should note that “agency” here means broad end-user access, not state agency. Countries without capacity to respond to AI-driven mass scams may asymmetrically bear the trade Altman calls acceptable.
Editorially, “accept bad things” becomes the week’s soundbite — comparable to Amodei’s “pace the frontier” — and will polarise policy panels through the quarter.
In short, Altman makes OpenAI’s bet explicit: more daylight from Anthropic, light-touch regulation, and tolerance for hacks/scams in exchange for massive benefits — a position now on the public record, with critics ready.
