The Takeaway

When people ask me what I've learned about leadership from reporting on tech companies for 20 years, I always say the same thing: that the most successful CEOs are more pragmatic than zealous. 

Tech CEOs are notorious for using visionary language and tying it to their personal beliefs about improving the world. They brandish mission statements about organizing all the world’s information and make us believe that their products are going to revolutionize the world for the benefit of humankind. 

But, in truth, they make far more pragmatic choices than their rhetoric often suggests.

Meta CEO Mark Zuckerberg is a classic example of this. I’ve watched him up close as he has pivoted his company from connecting people to building the metaverse to creating superintelligence—based on the moment. His views on content moderation have evolved based on politics—for which he’s received a lot of criticism, that’s fair in my opinion.

Elon Musk is another pragmatist. SpaceX’s mission was sending people to Mars until that seemed, well, too far fetched. Now, I’m not sure I could explain the company’s mission coherently. But Musk will think of something else grand to talk about publicly— and behind the scenes, he will keep pivoting.

While some might see this pragmatism a weakness, I think it is the only way you can build one of the world’s most valuable companies. 

Yesterday, when The Information broke a story that OpenAI’s CEO Sam Altman had agreed to a staggered release of his company’s latest model at the request of the federal government, my first reaction was that this was a sign of how Altman uses his pragmatism as a strength. (My second reaction was to ask what this meant for the future of tech regulation. But I will save that topic for another column.)

In Altman’s case, he has also been criticized for saying one thing and doing another. I’ve witnessed a little of that in my 20 years of knowing him.

But I think a more dominant aspect of his leadership style is his pragmatism. Altman is a leader who saw that a non-profit organization could never fund AI development and convinced his friend and investor Vinod Khosla to help him convert it into a for profit. Then he convinced Microsoft to do what was then one the world’s most advantageous compute deals, and so on and so on.

That’s textbook pragmatic leadership, and some may find it distasteful and opportunistic. It runs the risk of compromising trust. But if done well, it works and I understand why investors bet on it. 

One of the biggest questions hanging over AI now is what kind of leader Anthropic CEO Dario Amodei will be.

He’s a brilliant technologist, who knows more about the power of AI than probably anyone else.

But will Amodei temper his principles to adjust to the reality of the moment? Will he set aside some of his beliefs about the power of AI and how it should be rolled out for building relationships with governments, customers and more?

Right now many researchers celebrate him for not compromising. His standoff with the U.S. government, which began with a battle with the Department of War and has now extended to a conflict with other parts of the Trump administration over the release of its latest model Fable, has won him some big praise from many AI researchers and engineers.

Anthropic clawed back the release of Fable, after Amazon CEO Andy Jassy and others warned the administration of security vulnerabilities in the model, The Information was first to report. Anthropic downplayed the significance of the issues with the model, but the government effectively forced it to take Fable off the market. 

I understand why Amodei doesn’t want to get in the habit of making changes to its models at the behest of the government. That could be a slippery slope.

But so is the alternative, a world where its more pragmatic competitors strike compromises and beat it out the door. 

My gut tells me that even Amodei, who is directing a technology of never-before-seen power, will need to get in touch with his pragmatic side to win.