Mark Zuckerberg Admits Meta's AI Agent Development Falls Short of Expectations

Meta CEO Mark Zuckerberg has admitted that the pace of artificial intelligence (AI) agent development at his company has not progressed as quickly as executives had previously hoped.

As reported by TechCrunch on Thursday (2/6) local time, Zuckerberg shared this during an internal company meeting. AI agents are systems capable of understanding instructions, formulating plans, and completing tasks autonomously upon user command.

According to Zuckerberg, the development of AI agents at Meta has not yet shown the acceleration in progress that was previously projected.

This statement comes several months after Meta conducted layoffs affecting approximately 8,000 employees, or about 10 per cent of its workforce. In addition to these reductions, Meta also reassigned around 7,000 employees to various AI divisions, including a team known as Agent Transformation.

He noted that these steps were taken because company executives were concerned that Meta was not moving fast enough to adapt to the rapidly evolving technological industry landscape. Zuckerberg also mentioned that the anticipated benefits from the new AI-focused organisational structure have not yet fully materialised.

Nevertheless, he remains optimistic that the results of the company’s investments and reorganisations will begin to show a positive impact within the next three to six months.

Previous reports have described the environment within Meta’s newly formed AI teams as high-pressure, based on testimonies from several engineers reassigned to those divisions.

In recent years, Meta has become one of the most aggressive technology companies investing in the AI sector. The tech giant is reported to be allocating up to $145 billion for AI infrastructure throughout this year.