Foundation Labs vs Products
Had this shower thought today and wanted to share:
Every foundation lab needs a corresponding product.
Anthropic --> code (claude code + customers like cursor)
Grok --> X (social)
Meta --> Instagram (social)
OpenAI --> Enterprise tooling
Microsoft --> Enterprise tooling
When you're thinking about competition, you could say "both OpenAI and Anthropic are competitors because they're building AI models" but really, the competition is within categories, not between labs.
You nailed it. The competition is vertical, not horizontal, because differentiation at the foundation model layer is increasingly trivial. If everyone has access to the same ingredients, the winner is whoever delivers the best dining experience. That’s where the real job-to-be-done lies.