UpTrajectory Review

Sam Altman has pivoted from his September stance that AI development 'should be slower than it otherwise could be.' In a Politico Decoded podcast interview, the OpenAI CEO now frames his company's lighter-touch regulatory approach as a competitive differentiator from Anthropic, explicitly stating that society must accept 'some bad things happening'—hacks, scams, misuse—as the price of innovation. This reversal comes just weeks after a former OpenAI and Anthropic employee's resignation letter sparked industry-wide debate about responsible development, and after Anthropic CEO Dario Amodei published an essay advocating for 'pacing the frontier' to prioritize safety.

For small-business operators, this shift matters because OpenAI's products likely power your customer service chatbots, your marketing copy, your data analysis tools. When the CEO of the company behind ChatGPT publicly acknowledges that 'major hacks' and 'zero scams' are off the table as guarantees, he's signaling that your business absorbs the risk while OpenAI captures the upside. If you're building workflows on OpenAI's API or relying on ChatGPT Enterprise, Altman's candor is a prompt to audit your own exposure: What happens when the tool you're using gets weaponized against your customers, or hallucinates in ways that damage your reputation?

What's genuinely new here is not the position itself—tech leaders have long traded in 'move fast' rhetoric—but the explicitness. Altman isn't offering the usual bromide about balancing innovation and safety; he's making a utilitarian wager that 'orders of magnitude' more good than bad will emerge, and that resilience will develop organically as society absorbs the damage. We agree this is refreshingly honest, and it usefully clarifies the fault line: Anthropic is betting that frontier AI is dangerous enough to slow down, while OpenAI is betting that the diffusion of benefits outpaces the accumulation of harms. But we're skeptical of the framing that 'society figuring out resilience' is a passive process—small businesses don't have the capital to absorb the 'bad things' while they wait.

The second-order effects cut unevenly. Large enterprises with dedicated security teams and legal departments can internalize the risk Altman describes; a 20-person retailer cannot. This stance also shifts liability dynamics: if OpenAI's CEO is on record accepting that 'major hacks' and 'misuse' are inevitable, plaintiffs' lawyers will notice. Expect insurance markets to reprice AI-related cyber risk, and expect regulators—particularly in the EU, where the AI Act is already in force—to cite statements like this as evidence that voluntary commitments are insufficient. Altman may be speaking to investors and policymakers, but the downstream cost lands on the small operators who lack lobbyists.

Watch whether Anthropic leverages this in enterprise sales, positioning itself as the 'responsible' alternative for risk-averse buyers. More immediately, if your business depends on OpenAI tools, treat this as a signal to diversify: test Anthropic's Claude or open-source models for critical workflows, and revisit your vendor contracts for indemnification clauses. Altman has told you plainly that he won't trade innovation for your security. Believe him, and plan accordingly.

“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.” — Fast Company

Takeaway: Audit your AI vendor dependencies now—Altman just told you OpenAI won't guarantee protection from the inevitable downsides.