OpenAI is nearly free from its do-gooder shackles

By Dave Lee

Remember this?

“OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.”

That was the founding mission statement of OpenAI. It’s now quite out of date, so let’s do some light editing. Maybe something like:

“OpenAI is an artificial intelligence research company. Our goal is to advance digital intelligence to generate financial return.”

That is much clearer and far more accurate. On Wednesday, Reuters first reported that the company is set to announce a restructuring under which its nonprofit board will lose control over the company’s core business.

The new structure is designed to help make it more attractive to potential investors. Recall that it was OpenAI’s nonprofit board that moved dramatically last November to oust Chief Executive Officer Sam Altman, leaving its biggest backer at the time, Microsoft Corp., wondering if its $10 billion AI strategy was about to turn to dust. A governance change should stop that from happening again.

The move to reduce the board’s control comes as, any day now, OpenAI is expected to confirm that it has received the biggest venture capital investment ever: a $6.5 billion (or more!) injection valuing the company at around $150 billion. Those investors aren’t doing it for the good of humanity; they’re doing it for the good of their funds and partners. Unlike investments made under OpenAI’s original structure, the Wall Street Journal reported, “those who put money into the current round wouldn’t have a cap on the profits they can earn.”

Still, before any of them have any hope of being paid back, OpenAI must turn around an eye-watering deficit between its revenue ($3.6 billion annually, according to reports) and its costs (more than $5 billon). Even if the company decides to restructure as a public benefit corporation, as Bloomberg has reported is an option, those sums don’t change. If you think OpenAI can possibly operate “free from financial obligations,” you’re hallucinating. That dream is over. (The cynic in me is pretty sure it never existed.)

It’s not the only transformation at OpenAI. From its original 11-member founding team, just two remain: Altman and Polish computer scientist Wojciech Zaremba. (Greg Brockman announced earlier this year he was going on “sabbatical” until the end of the year.)

On Wednesday came news of another prominent departure: Mira Murati, OpenAI’s chief technology officer, who wasn’t among those founders but was an influential executive who stepped in briefly as CEO when Altman was pushed out.

Murati gained media prominence after poorly attempting to dodge a question from the Journal about whether OpenAI had been scraping videos from YouTube to build its models. She said on Wednesday that she was “stepping away because I want to create the time and space to do my own exploration.” (Not long after she shared her note, two more departures, from OpenAI’s research team, were announced.)

We don’t yet know Murati’s plans, but her unexpected departure shares some of the hallmarks of other senior executives who went on to join or launch competing AI companies. Ilya Sutskever, who announced he was leaving OpenAI in May, recently raised $1 billion for his new company, Safe Superintelligence, which pledges to work solely on safety and has “no near-term intention of selling AI products or services.” (With “techno-optimists” Andreesen Horowitz among the backers, let’s see how long that promise holds up. There’s that cynic in me again.)

Sutskever’s exit was part of a pattern of high-profile executives who left OpenAI with departing messages (or future actions) that suggested concern about the direction of OpenAI under Altman. According to the Journal, Murati had previously described Altman’s leadership style as “psychologically abusive.” Finding it impossible to force change at OpenAI from within — Sutskever had cast the deciding vote to try to force out Altman — they have ended up as the ones departing. Altman has seemingly neutralized any opposition within the company.

This is all to say after a turbulent year, OpenAI’s metamorphosis is almost complete. The course change, which was set in rapid motion by the failed ouster of Altman, will be fully realized once the new structure is confirmed and that gigantic VC check clears. It will mean the company that had admirable checks and balances, and a CEO without equity in his own creation, now looks a lot more like the other tech giants it competes against. Maybe that’s the only way it can keep up.

Crucially, OpenAI 2.0, unlike its original incarnation, is very much Altman’s company. He will, reports suggest, finally be given equity valued at around $10 billion. No longer constrained by hollow pledges of good before profits, and with new financial incentives of his own, we’ll see whether the concern of those who exited was justified.