I agree with most of the premises in this article. It reminds me of one of my favorite seminal a16z articles: https://a16z.com/2020/02/16/the-new-business-of-a

I agree with most of the premises in this article. It reminds me of one of my favorite seminal a16z articles: https://a16z.com/2020/02/16/the-new-business-of-ai-and-how-its-different-from-traditional-software/ https://a16z.com/2020/02/16/the-new-business-of-ai-and-how-i... Basically, a16z argue that AI businesses by default are closer to consultingware than software companies because their data hungry nature works to impede rather than enable unit economics. I find this to be broadly accurate. Many of the existential challenges facing a would be AI implementation include data unit economics and domain validation -- the 0-1 product problems, not the 1-1000 problems.