AI agent research and development is hindered. Zuckerberg admits that the effect of Meta's transformation and restructuring is not as good as expected.

According to Reuters, technology giant Meta CEO Mark Zuckerberg admitted at an internal company meeting on Thursday that Meta's research and development progress in the field of artificial intelligence agents (AI Agents) has not been accelerated as expected in the past four months, and the development speed has been significantly slower than the company's previous plan.

According to an internal meeting recording obtained by Reuters, Zuckerberg told employees at the meeting that in addition to the hindered progress in AI research and development, the company's previous large-scale organizational restructuring and layoffs in order to fully transform to AI were not carried out "cleanly" enough. He admitted that Meta’s many ideas and initiatives on the new organizational structure “have not yet achieved the expected results.”

Earlier this year, due to Meta's senior management concerns that the company was not advancing fast enough in the field of AI, Meta carried out drastic business adjustments in May, including cutting 10% of its employees globally and redeploying about 7,000 employees to new positions related to AI workflow and model training. However, this radical functional shift not only triggered internal concerns about the surge in workload, but also caused a certain impact on corporate culture. Zuckerberg previously admitted in an internal letter that the company did make some mistakes during such a complex transformation.

Although he admitted that the research and development of AI agents in the short term has not achieved explosive growth expectations, and the dividends of restructuring have not yet appeared, Zuckerberg emphasized at the conference that from the perspective of long-term development trends, Meta's strategic direction of firm transformation is still correct and unshakable. He said that Meta will continue to devote resources to AI infrastructure, top talents and cutting-edge models, and expects that in the next three to six months, employees and users will gradually see the actual benefits brought by these AI layouts.