Meta accelerates agent talent acqui-hire spree

This looks less like panic buying and more like Meta trying to own the agent stack before the category hardens. The bet is smart if agents become the next interface for work and commerce, but the integration bill could get ugly fast.

• –Manus gives Meta a consumer-ready agent product with real subscribers, which is a faster route into Meta AI, WhatsApp, and Instagram than building a rival from scratch.
• –Moltbook and Dreamer both point to the same UX and ecosystem thesis, and Dreamer's short beta window shows Meta is willing to buy momentum early.
• –Putting Wang in charge makes Superintelligence Labs a central command point, but it also concentrates execution risk if too many teams are folded together too quickly.
• –The "4th startup" framing is a little squishy, because Scale was a strategic investment plus talent transfer rather than a classic acquisition.
• –Open-source Llama still matters, but Meta is clearly betting the moat will come from the application layer, not the base model alone.