I'm curious about the current AI startup landscape and how many companies are essentially just wrapping foundation model APIs (OpenAI, Anthropic, etc.) with a UI layer versus doing more substantial technical work.
Some questions I'm interested in:
- How can you tell if a company is primarily using foundation model APIs?
- What percentage of AI startups fall into this category?
- Are there examples of companies doing this particularly well or poorly?
- What constitutes legitimate value-add on top of foundation models?
