Elon Musk is a genius. He has a remarkable ability to defy conventional wisdom and do what others consider impossible. We should be grateful to him for revolutionizing at least two different important industries (Tesla/automobiles and SpaceX/space) and for tackling many others (Starlink/telecommunications, Hyperloop/transportation).

However, being a genius does not mean being omniscient/omnipotent. Elon Musk’s achievements depend on a very specific formula, and his ventures outside this formula are unlikely to succeed.

The Elon Musk formula is simple: Apply first-principles engineering to an inspiring quest that no one else will tackle because it is believed to be impossible.

Tesla follows this formula. People didn’t think electric cars would work; Elon realized he could use a massive number of laptop batteries, and ride someone else’s cost curve down. Thousands of great engineers signed up to be a part of Tesla’s quest.

SpaceX follows this formula. People didn’t think private spaceflight was economically possible; Elon realized that he could use modern technology to build reusable rockets. Thousands of great engineers signed up to be a part of SpaceX’s quest.

This formula allows Elon to get away with being one of the worst people managers in history. I have spoken with quite a number of people who worked for Elon at one of his companies. They all tell me the same thing: “Working for Elon was amazing. Together, we accomplished more than I thought humanly possible. There is nothing in the world that would ever convince me to work for him again.”

Now consider some of Elon’s new projects. First-principles engineering won’t fix Twitter. Fixing Twitter isn’t an inspiring quest, in part because there are so many other more successful social networks. There is no reason for great engineers to work for Twitter unless they are ideologues, which severely limits both the quality and quantity of his technology team.

Or consider Elon’s plan to start an OpenAI competitor. Why would anyone join his effort when every other major technology company is already many years ahead? There is no reason for great engineers to work for “Elon.ai” unless they are the equivalent of cult followers, and as I’ve commented before, I can’t think of any cults that are financing their operations by selling market-leading software.

Elon is a big fan of the character of Tony Stark. He even filmed a cameo in Iron Man 2. But Tony Stark is a cartoon character. Omnidisciplinary scientists do not exist in real life. Even a genius as brilliant as Elon Musk can’t do everything. Let’s hope that he eventually recognizes the true secret of his success and achieves his goal of colonizing Mars and the rest of space.