GPT-4, race, and applications

GPT-4 came out yesterday and overshadowed announcements, each of which would have been bombshell news otherwise:

  • Anthropic AI announcing their ChatGPT-like API -- likely the strongest competitor to OpenAI today (only waitlist)
  • Google announcing PaLM API (currently it's just marketing -- no public access yet)
  • Adept AI announcing their $350M round (they build foundation models to interact with anything on your computer)

The race is clearly on for building the core tech of ~general AI.

What's more interesting to me, though, are applications. Based on my last Linkedin poll (56 votes, definitely biased towards AI early adopters):

  • one half have a Definite optimistic view: they think AI will affect their industry & company, and they know how
  • one third have an indefinite optimistic view: they think AI will affect them but they don't know how
  • the rest think AI won't affect them much

(I am of course twisting the term "optimistic" here a bit away from Peter Thiel's original formulation.)

I have already been affected. Personally, I'm an almost daily active user of ChatGPT (use cases: humor, code, learning, writing, research, data conversion, etc) and have built many experiments on top.

The stickiest one I've built for myself is a command-line helper, where I can describe in free text what I want to do, and get back a recommended command to run. The most valuable case was when I figured out a complex set of ffmpeg arguments with just a few tries.

What I would love to hear more of is: how has everyone else used these models, and what have you found valuable enough to keep using several times or even daily? (I'd love to hear from you if you're reading this -- shoot me an email!)