I Don't Think Sam Altman Lies

By now, most of you already know that The New Yorker published a very long report titled “Can Sam Altman Be Trusted?”. (If not, it’s an entertaining and very well reported story.) It is one of the most talked about stories ever in Silicon Valley, partly because it does such a good job of being balanced that it’s like holding a mirror up to the face of the tech industry.

I’ve read the story and read and listened to a lot of the industry chatter about the article and about Sam Altman, whom I’ve never met. (Hard Fork interviewed the authors, which was an interesting deep dive. They got close to my thesis, but were too polite to go all the way.)

My thesis: Whatever his other attributes, Sam Altman is one of the worst CEOs I’ve watched. I’ve been following tech (and non-tech) companies for nearly five decades now, and have watched dozens of CEOs, good and bad. I can spend hours talking about what makes a great CEO, or not. Sam is likely a perfectly honest and well-intentioned individual. And clearly very, very smart.

If I was really on my game, I would tell him to resign as CEO for the good of the company. But, since he got fired the wrong way the first time and the company’s current board isn’t positioned to challenge him, there isn’t likely anyone besides me who would tell him that to his face. I’ve never met Mr. Altman, and he’s not likely to come to the same conclusion on his own, since he doesn't appear to be the kind of guy who reflects on his strengths and weaknesses.

I think of OpenAI as the Twitter of Generative AI. Twitter was a company founded by mistake. It churned through CEOs and never really found its place until Elon Musk came along and bought it for $44B. OpenAI is a very different beast than Twitter; the company has raised gobs of money, $180B, and is now valued at $850B — before it tries to go public.

Twitter did manage to go public and was valued at $44B before Musk bought it and took it private. It looks like Musk is the first CEO to wrestle it to the ground before turning it into the social media component of his space AI company. Being a public company generally speaking requires a highly disciplined CEO who has the ability to steer the ship with a steady hand and a strategy that is consistent over time. This is the problem with Sam Altman: He doesn’t know how to do that and hasn’t been trained for it.

Indeed, the The New Yorker article makes that explicit. He appears to hire really well and then not know how to work with his own executives. He hasn’t ever built a high-performance executive team. He’s a great investor and pitch man, but that doesn’t sit well with institutional investors. And he does not have the actual accomplishments of an Elon Musk (Tesla, SpaceX, Starlink, etc.) and hasn’t created billions of dollars of permanent value in a prior life. Put simply, even I wouldn’t buy stock in a publicly-held OpenAI. (But see my last newsletter for why that doesn’t matter; no one should ever follow me investing in public companies.)

OpenAI has been swinging wildly from one tactic to another with no consistency. Canceling Sora as a “side bet” one week and then buying TBPN the next week is a pretty good example of a lack of strategy, something most people define as “knowing when to say no”. This lack of strategy is compared, of course, to Anthropic, which has been totally consistent: focus on the enterprise, figure out how to turn coding into a competitive weapon, and double down on its ability to matter. Even the U.S. Department of Defense will have to eventually acknowledge that Mythos matters to our national security way more than Pete Hegseth’s ego.

So, @sama, what do you think about finding a new job and letting someone else take the lead? I wonder if you could recruit Dara Khosrowshahi to your board and actually listen to him since he’s the guy that figured out how to fix Uber after Travis Kalanick made a mess of it.