Zuckerberg’s Bets on Automation Go Up in Smoke

Mark Zuckerberg’s high-stakes gamble to remake Meta around artificial intelligence has hit a speed bump. The tech mogul’s plan to revolutionize the way the company operates by shrinking teams and replacing them with AI agents that could absorb most of the daily work has been put into question.

At its core, Project OT aimed to eliminate middle management layers and create small groups of humans overseeing virtual workers. These “talent-dense” pods would consist of three to five engineers, designers, and researchers working together under one generic title, “builder,” with no formal authority.

However, Meta’s employees had other ideas. When the company started tracking keystrokes and mouse clicks on US laptops to teach AI agents how humans use computers, over 1,000 employees signed a petition against it, and internal posts filled with elephant images to mark the layoffs nobody would discuss. The backlash suggests that even in a tech-savvy industry like Silicon Valley, people still value their jobs.

The technology itself also struggled to live up to its promise. Despite significant code changes, actual user-facing improvements were modest. Major technical and security incidents climbed 40%, and time spent addressing them increased by 70%. It’s clear that AI is not yet ready to take the reins, at least not in Meta’s case.

Now, as the company rebuilds its management layer, it’s offering a voluntary route back into management for those who were moved into Applied AI this year. This reversal of fortunes marks a stark departure from what was once touted as a brave new world where humans would oversee virtual workers.

Zuckerberg has conceded that the reorganization was mistimed, and agent technology hasn’t moved as fast as he expected. The bet was that AI could run the work with fewer people watching over it, but Meta is discovering how many people are still needed to keep things running smoothly.

The implications of this failure go beyond Meta’s walls. As companies continue to invest heavily in AI infrastructure – at least $130 billion earmarked for Meta this year – they may be overlooking the value of human workers. The tech industry has long been plagued by concerns over job displacement, and Zuckerberg’s failed experiment serves as a reminder that automation is not yet a silver bullet.

In fact, this setback might be a blessing in disguise. By slowing down its AI transformation, Meta can take a step back and assess what really works in the industry. Perhaps it’s time to rethink the role of humans in tech – not just as overseers, but also as creative problem-solvers who bring value to the table.

The company’s decision to rebuild its management layer rather than fully embracing automation suggests that human intuition and experience still have a place in the tech industry. As the old saying goes, “garbage in, garbage out” – and it seems that Meta’s AI still needs some human touch to function properly.

As the company moves forward, one thing is clear: Zuckerberg’s bet on automation has gone up in smoke. The future of work remains uncertain, with companies continuing to invest heavily in AI while also reevaluating their reliance on automation. Only time will tell whether this setback marks a turning point or simply a minor detour.