Reading between the lines of the Llamacon Keynote

[Hot take:]

The first Llamacon keynote just wrapped seconds ago, and I feel like I'm getting a sense of Meta's AI strategy for the first time. They didn't say it directly, but you could hear it between the lines.

Many had speculated Zuckerberg was pursuing a "commoditize your competitors" approach, out of fear of being trapped as an app within yet another company's platform again. I don't think that's it.

If AI and AR represent an entirely new computing paradigm, that new paradigm will require a new OS. And that new OS will require a host of standard utilities, like GNU utilities were to Linux.

  • Small, fine-tuned models
  • Large, stock models
  • Realtime voice models
  • 3D understanding models
  • Image segmentation models
  • Scene generation models

Collectively, that sounds a lot like the stdlib for a completely different platform of AI/AR computing.

The insistence that all Llama derivatives be prefixed with llama- feels telling:

The last 40 years we've all been building atop GNU/Linux. I think in 5 years, Meta wants us to all be builting atop LLama/Something.

© 2025 Edward Benson. All rights reserved.