Zuckerberg Admits Meta's AI 'Hasn't Really Accelerated' After Cutting 8,000 Jobs

My take: Meta went all in on AI agents: it cut 8,000 jobs (10% of its workforce), reassigned 7,000 more employees to AI-focused teams, and committed up to $145 billion in AI infrastructure for this year. This week, Zuckerberg told staff directly that the AI agent development effort "hasn't really accelerated in the way that we expected."

This reveals something important: AI adoption at enterprise scale is harder, slower, and more expensive than the headlines suggest. Not because AI does not work, but because integrating it into real processes, with real people and legacy systems, takes more than a reorg and a massive budget.

For any company, large or small, the lesson is clear: AI is not a switch you flip on — it's a capability you build over time, with experiments and with people who actually know how to use it. Is your team developing those skills now, or waiting until the "product is ready"?