Altman claims ChatGPT has more users in Texas alone than Claude has across the entire US

OpenAI CEO Sam Altman claims ChatGPT has as many users in Texas alone as Claude has across the entire United States.

Altman dropped this statistic while responding to Anthropic's recent promo videos that mock ads in ChatGPT. The jab is particularly pointed given that Anthropic had at the same time said it wouldn't run ads for Claude.

According to Altman, the commercials are entertaining but misleading. OpenAI has explicitly rejected the kind of advertising Anthropic depicts in its spots: subtle influence woven into personal, confidential chat conversations. He also accuses Anthropic of elitism: OpenAI wants to bring AI to billions of people who can't afford subscriptions, while Anthropic primarily caters to wealthy customers with an expensive product.

Altman's promises face a credibility test

Altman might be right with his assessment, but there's some irony here. The dystopian scenario Anthropic sketches is one Altman himself once warned about, back when it fit his anti-Google messaging. And with OpenAI planning an IPO later this year, the pressure to monetize could push the company toward more persuasive and intrusive forms of advertising. For publicly traded companies, ads have a way of creeping in and gaining influence when growth targets need to be hit. It's a relatively easy growth lever to measure and adjust.

Why is persuasion in chat such a big deal? Back in 2023, Altman warned that AI would achieve "superhuman persuasion" long before general intelligence and that this could lead to "very strange outcomes." And besides Altman's warnings—and being responsible for ChatGPT—reality has since proven him right: lawsuits over chatbot-related suicides, documented cases of users developing intense emotional dependencies, and research showing AI conversations are more persuasive than static messages.

Add advertising incentives to that mix, and what Anthropic mocks in its videos starts to look less like satire and more like a warning. Spelling it out seems fair, and if it helps keep Altman and OpenAI honest, all the better.