It can write you a function/method/script if you give it an outline of what you need but it's awful at anything remotely complex and everything it spits out needs to be understood by somebody so it can be checked.
It can save a developer time but it only replaces people in that one person can be more productive which is absolutely nothing new.
The reason for that is because it's not AI, it's ML. It doesn't understand anything.
Those languages came about during the boom of websites and apps, however.
And more code was being written, rewritten and maintained by coders--now, instead, it's looking like it'll be generated, tested, analysed and regenerated.
Almost every company I've worked at have more ideas for projects than coders. The constraint has been how much they're willing to spend on R&D and not how much work they can think up.
If one developer can all of a sudden do the work of four, I think leadership is just going expect higher output from the same number of people.
This is the only comment that has made me less worried about AI.
If AI gets to the point where an end user can easily describe the app or website they want, however, then it won't just be the programmers who are in trouble.
It can write you a function/method/script if you give it an outline of what you need but it's awful at anything remotely complex and everything it spits out needs to be understood by somebody so it can be checked.
It can save a developer time but it only replaces people in that one person can be more productive which is absolutely nothing new.
The reason for that is because it's not AI, it's ML. It doesn't understand anything.