I'm not Sam but I wanted to talk about this anyway. I'm working on a side project that deals in the AI space (would like for it to become a business; we'll see how things go when I launch).

Anyway, Chris's article confuses accuracy with capability and is very strictly talking about synchronous AI agents. But I think there is a ton of space for AI related start-ups that don't fall into the trap he describes.

I'm working on a side project that doesn't work in most of the ways he describes (made it at the Launch Hackathon a few weeks ago if you're interested; website should be launched this weekend but this was my submission: http://devpost.com/software/sim).