4 ms·
> any layperson can say "hey AI slave, write me a program that..."? I think this is one of the big challenges. From my limited experience, a person without exp
by _benj 4y ago
> any layperson can say "hey AI slave, write me a program that..."?
I think this is one of the big challenges. From my limited experience, a person without expertise being able to describe accurately enough what a program should do is often a challenge. Further more it is often after an initial implementation that even engineers are able to understand the full requirements of a system. It's possible that generating yet-another-crud-app wouldn't be any challenge to GPT but then at that point there could be a market saturation of CRUD applications (is that not the case already?) and value might shift to apps that are able to do more than just CRUD.
Then again, predicting the future is tricky business when we are so notorious at not being able to predict what can we do during the next sprint! :-D
- skydhash 4y agoDescribing accurately what a program to do is already possible with a programming language. But it requires logic and an understanding on what a program can do and how it does it. People that want AI either wants a way to help deduce the logic needed to solve a problem or want to automate the boring part of the implementation. The first can already be done by researching and the second by creating and using frameworks. Unless we find a way to add actual reasoning and creativity to AI, I wouldn’t feel threatened. Because coding do involve the two. As well as writing and painting. So there will always be a human at the helm. But what AI is becoming good at is finding patterns and reproducing outlined steps. And if you job can be resumed to these, then you’re in danger.