1. Home
2. News
3. OpenAI

Altman quotes Amodei's 2024 essay without naming him, and says OpenAI is against the idea

He says keeping people in charge should ultimately constrain AI progress, and doesn't say how.

By Think Facility · November 7, 2026 · OpenAI AI safety Save

On the evening of November 3, 2026, Sam Altman posted three sentences on X about who should run the future. People should always be in control of it, he wrote, and OpenAI is against the idea that people could end up doing whatever AI systems decide is worth rewarding in humans.

The words he put in quotation marks are Dario Amodei's. They come from Machines of Loving Grace, the essay Anthropic's CEO published in October 2024. Altman's post doesn't name him or link to it.

What Amodei wrote

The line sits in the essay's section on work and meaning. Amodei writes there that once AI can do most of the work, the economy will probably need some new arrangement nobody has pictured well yet, and that a large universal basic income might be only a small part of it. Then he lists options.

Amodei offered it as one possibility among several. In the same passage he says humans might stay economically valuable in ways the usual models don't expect, and that you couldn't tell which option works without a lot of trial and error.

Altman's third sentence goes further than disagreeing with one idea. He calls keeping people in control a requirement, and says it should ultimately constrain AI progress. From the person running OpenAI, that's a statement that progress should give way to control. The post doesn't say what the constraint would be, or who'd set it.

What he said on a podcast

Altman made a similar point at more length in an episode of Raj Shamani's Figuring Out podcast, posted on YouTube with the title Sam Altman On AI Revolution, Superintelligence & Future Of Humanity. He named two big risks.

The first is a loss of control, where people hand the future to systems far more capable than they are, on purpose or by accident, and can't get it back. The second is one company or country getting far too powerful next to everyone else, which he called very unstable.

He also talked about his rivals. Many people, in his telling, end up trusting only themselves to build this technology because the stakes feel so high. I'd read that next to the post, which takes on a competitor's idea without naming the competitor.

What it means for people's work

The two CEOs are arguing about an economy that doesn't exist yet. A small version of the question can already be measured, though, and Anthropic has measured it on its own staff.

In its report on employment at the company, Anthropic says junior technical hires fell from 27% of its technical hiring to 9%. In an April 2026 survey, when staff were asked how AI had changed their day-to-day work, 20% described their role shifting from doing the work to managing Claudes.

That's one company's engineers, at the company building the models. Still, it's the nearest measured thing to the arrangement Altman says he's against: people whose job is shaped by what AI systems are doing. Neither CEO's words deal with numbers like these. Altman's post is about who stays in charge, and Amodei's passage is about how the money gets shared.

What hasn't been said

Altman didn't say whether OpenAI would accept a limit on its progress set from outside the company. We didn't find a reply from Amodei to the post.

Altman's post went up at 8:40 pm Eastern on November 3, and it doesn't link to any OpenAI policy.