[JamesBarney]

Fundamentally it's about risk. If your startup is writing OCR that out competes tesseract or automatically diagnoses melanoma that has very different risks than building a B2B product to allow optometrists to manage their clients.

In general though I think most startup's have more go-to-market risk than technical risk. If you can out compete tesseract in OCR you're not going to fail because of market. And if you're Salesforce you're not go to fail because you can't build software to track leads. You're going to failed because you couldn't find enough users.

This is why investors spend so much more time looking at adoption rates, and so little time looking at the backing infrastructure.