The deal closed in roughly ten days. That's the same speed Meta used for the $2B Manus acquisition in December. Now, with the execuhire of Dreamer — the team behind consumer AI agent Sidekick — the pattern is unmistakable: Meta Superintelligence Lab is running a scorched-earth talent acquisition strategy in the consumer agent space, and money is no object.
The timing reveals the urgency. Dreamer appeared on the Latent Space podcast on March 13th. Eleven days later, Nat Friedman and the MSL team announced the execuhire. No extended due diligence. No months-long negotiations. The team had barely finished describing Sidekick's architecture on air before they were Team Zuck.
"If you can't tell from the pod, we were immediately in love with the tech and the polish," Latent Space wrote. "But it was always going to be a long slog to build any consumer AI business, and it is a very nice thing indeed to have Team Zuck on your side."
What Meta is buying isn't primarily the Sidekick product — impressive as it is. The $2B Manus deal already brought strong agent technology into the fold. What Meta needs is the *experience*: the understanding of how consumers actually interact with AI agents in daily life, how to build an agent ecosystem, and how to make a personal superintelligence feel intuitive rather than industrial.
Sidekick's architecture reflects this know-how. It's built as an "agent-of-agents" — a meta-layer that coordinates multiple specialized AI capabilities to serve individual user goals. This is precisely the vision Mark Zuckerberg articulated nine months ago in his "Personal Superintelligence" manifesto: "Personal superintelligence that knows us deeply, understands our goals, and can help us achieve them will be by far the most useful."
Dreamer's team built toward that vision in the consumer market. Manus brought distribution and technical depth. Together, they represent a talent concentration that no other company currently matches — unless you count OpenAI or Anthropic, both of which are pursuing similar strategies with their own agent initiatives.
The execuhire model itself is notable. Unlike a traditional acquisition, "execuhire" implies licensing plus hiring without full corporate absorption. This suggests Meta wanted the people and their knowledge, not necessarily every asset or liability. For a company racing to build consumer-facing AI at scale, speed trumps thoroughness. You buy the team; you let them build.
The competitive implications are stark. If Meta now fields Manus plus Dreamer — two of the most-discussed consumer agent teams of the past year — it has effectively locked out competitors from a finite pool of proven talent. Google, Apple, and smaller AI startups are watching the same playbook and drawing the same conclusion: the consumer AI agent race is increasingly a talent war fought with acquisition checks, and Meta just raised the stakes.