Mark Zuckerberg said out loud what many in the industry have been thinking quietly: AI agents are developing slower than expected. At an internal town hall on July 2, he told staff that development “has not accelerated in the way we expected” over the last four months.
The context
Meta cut about 8,000 jobs earlier this year — roughly 10 percent of its corporate workforce — and reassigned another 7,000 employees to AI groups, including a unit called “Agent Transformation.” Zuckerberg acknowledged that the restructuring “wasn’t as clean as it could have been” and that the bets on the new structure “haven’t come to fruition yet.”
Still optimistic
Despite the candor, Zuckerberg is holding the line: he expects the investments to pay off within three to six months. Meta raised its capital expenditure forecast in April to between $125 billion and $145 billion. That’s an enormous commitment — and Zuckerberg needs to deliver.
My take
The honesty is refreshing. In an industry that loves words like “revolutionary” and “groundbreaking,” the CEO of one of the biggest tech companies is saying: our agents don’t work as well as we want yet. It confirms what many of us experience in daily use — AI agents are impressive in demos but bumpy in practice. Claude Code and Cowork are making real progress, but there’s still plenty of work left. The real question isn’t whether AI agents are coming — it’s how long that last mile actually takes.
Sources: