Gshaheen

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?