3 ms·

All of those examples are still just repeatable processes. The machines aren't making decisions independently and then executing complex tasks based on their ju

All of those examples are still just repeatable processes. The machines aren't making decisions independently and then executing complex tasks based on their judgment. Not saying we'll never get there with AI...but it's a fundamentally different problem space.

bognition 3y ago
A huge percentage of the software that is getting built today is little more than a CRUD app with some sort of domain specific data model.
- sublinear 3y ago
  Yes, but app maintenance and integration is what has always kept devs employed. Anyone can arbitrarily define what a v1.0.0 release looks like. It doesn't mean anything of value was really created, nor can we cut humans out when requirements need to be added.
- rco8786 3y ago
  Tbh I completely disagree with this statement. Even extremely simple CRUD has unique data modeling, validation, storage and scaling, deployment, UI/UX concerns, etc A very, very small percentage of software falls into the category of “simple form on top of a database table”. And even when it does, it doesn’t tend to stay that simple unless it’s just kinda not that valuable anyway.
famouswaffles 3y ago
LLMs can make decisions independently and execute tasks based on their judgment. I'm not putting a GPT-4 agent in charge of much of course, but that’s a competency problem, not a problem space one. 4 is a much more competent agent than 3.5. Presumably, 5 will be another jump from 4. You can't have these competency improvements and expect humans to always be necessary. And "just so competent, a human is needed" sounds like a real tangible line but is actually entirely arbitrary and no harder than any other line to cross.