What Sam Altman Is Hiding: Can We Trust People Who Promise to Change the World?

These words belong to Simon Rich, the screenwriter of "Artificial," a film about OpenAI, Sam Altman, Ilya Sutskever and Altman's firing. According to Alexander, the premiere took place on October 5 in New York; at the end of December the film opens in some US theaters, and goes wide from 2027. Rich wanted to draw attention to where the creation of AI is heading; whether it worked, Alexander does not know.

The film's main question is how far we can trust people who promise to solve humanity's problems while gaining more and more power. Most people do not want to think about it in relation to themselves: they care more about how to choose a profession and earn money. And the race remains — with AI people do fifteen tasks instead of three, the result is the same, and now they want fifty.

At the end of September OpenAI named a personal AGI for everyone among its three points of the future — not a tool for big companies and not a chat. Getting an AGI, Alexander says, is like getting governments, economies and businesses working for your life, like the difference between a single café and Starbucks. And your wife or husband, parents and children will each have such an AGI too.

Sam Altman believes we are already in the technological singularity. In an AI-generated video on X he comes to Jensen Huang, sitting on a "Game of Thrones"-style throne, saying "We seem to have reached AGI," and hears: "You've been saying that for three years" — while the prepared cake has rotted in the meantime. AGI, in effect, gets released every year or every month.

If AGI has been created, what should change in the world, and do its creators understand what they are doing? Alexander considers Elon Musk the most sane-minded — although he has a huge number of questions for him: Musk at least throws big trends forward. The heads of OpenAI and Anthropic, and even more so Google and Meta, in his view reason in micro-zones — not because they are unintelligent, but because of what they are willing to say openly.

The film's first level is a scientific dream colliding with the will to win — that is, with power. A month ago Altman said AI must be restrained, and then that humanity will have to accept bad things, without saying which. And Alexander's partner reminded him: all of today's AI is "the simplest, dumbest artificial intelligence there will be in the future."

In the film, Sutskever believes in the technology's enormous possibilities but gradually begins to see the danger, while Altman ties the company's future ever more closely to his own leadership. Hence the conflict, after which Mira Murati briefly led the company, and everyone said OpenAI was done and Zuckerberg had poached the team.

Today OpenAI has the most weekly and daily users and the most time spent in the system: at DevDay Altman named, as Alexander remembers it, about 1.2–1.25 billion. That is serious for a company that only has ChatGPT and Codex — no search, no email, no file storage — while Microsoft, for all the big partnership, holds back its use. Apple let ChatGPT in, then switched to Gemini, but Siri searches very badly.

Building serious systems takes not only money, technology and great teams but also what happens inside. What would OpenAI be if Sutskever and Murati had stayed, if Anthropic had not split off, if Musk had integrated OpenAI into SpaceX and Tesla? And at DevDay Alexander had a question: why does Altman read everything from a monitor and not talk to developers openly?

The film's second level is that talk about safety can also serve business interests. A warning about a system's incredible danger also underlines its incredible power. We are constantly told AI must be restricted, while the system never stops developing: "No one has ever stopped anything yet" — and its creators' predictions keep being broken.

Elon Musk speaks of a probability in the tens of percent that AI will lead to the collapse of humanity. Such a system, Alexander believes, could be launched at a probability of one in a million or one in a thousand, but not tens of percent. The Chinese call this narrative the Americans' way of slowing everyone else down, and Italy once blocked ChatGPT — yet today models have penetrated everywhere.

The film's third level is truth becoming a tool of control: for Sutskever it is a principle, for Altman something that adapts to circumstances. Everyone who resisted ends up buying compute anyway. Ilnar once said Anthropic had "started to back down": it promised not to work with certain countries and was already taking their money.

Anthropic's Mythos-class systems, according to Alexander, started making the first serious hacks in the world, and access was given only to certain companies by agreement with the US government. Ordinary people on free and cheap plans are nowhere near having access to serious models, while in the five-hundred-dollar ChatGPT Pro even the Extra High mode answers instantly — which is why paid models are worth investing in.

The fourth level is the double meaning of the title: "artificial" can also describe Altman's carefully built public image, his persuasiveness and his ability to build consensus in society. People trust those who stand behind the money, but five years ago these people had no idea where we would be and keep changing their timelines — so we have to figure out the truth, the causes and the effects ourselves.

In the US they decided to drop the word "artificial," and that, Alexander says, is no accident: Musk has long talked about superintelligence. We are far from the AI of films like Oblivion, but calling AI dumb code is wrong — it is a huge infrastructure and part of society's consciousness. It has no human spiritual-and-soul system, but it is more serious than an operating system: an OS never became part of society, and Apple can be replaced.