Mark Zuckerberg hired Alexandr Wang a year ago to build a new AI model. The model does not work. Now Zuckerberg has to explain why he spent billions on something that doesn't work.
This is not a failure of vision. This is not a pivot. This is a man who hired someone to solve a problem and the problem did not get solved. Wang runs a company that labels training data. Zuckerberg thought labeling data better would make AI better. Turns out you also need the AI part to work.
The results are underwhelming. That's the word they used. Underwhelming. Not catastrophic. Not a disaster. Underwhelming. Like a birthday party where someone brought store-brand chips. Except the chips cost twelve billion dollars and the party is a shareholder meeting.
Retail traders saw Meta's AI spending and thought it meant Meta was winning the AI race. They bought the stock. They told their friends to buy the stock. They made TikToks about buying the stock. What they did not do is ask whether the AI actually worked. Because that would require reading past the headline.
Zuckerberg is now on a sales tour. He's telling investors the AI will work eventually. He's showing slides. He's talking about long-term value creation. He's doing everything except showing them the AI doing something useful. Because if the AI did something useful, he wouldn't need to do a sales tour.
Wang got his job because he convinced Zuckerberg that better training data solves everything. Zuckerberg believed him. Zuckerberg gave him money and engineers and a year. Wang gave him underwhelming results. That's the transaction. Money for underwhelming results. Most people call that a bad deal, but in Silicon Valley they call it building for the future.
The future better get here fast because the shareholders are reading the present tense.
Photo by Julio Lopez on Unsplash
Leave a Comment