I fully agree. I have been using Copilot for a while and the errors it makes in more complex domains (algorithms and data structures more complex than 'write me a quicksort') are just stupendous. Even at a more basic level, it regularly makes Rust syntax errors that my 10yo daughter has no issues spotting.
It has been really awesome for writing boilerplate code though. Especially with in-context learning, writing one example and then letting the model extrapolate it to other methods, functions, etc. works great and saves me a lot of time.
It's hard to predict the future. But the current best models feel more like a much better IntelliSense (though API hallucination is a serious issue) than something that is going to replace good programmers anytime soon (unless your task is writing boilerplate code).
Nevertheless we talk about the future.
A lot of things seem to have a great potential and then peak a lot lower. AI producing very impressive but very short videos gives us a sense that it has the potential to produce whole, coherent films that tell a story.
The drivers have always been convenience and cost. AI is the ultimate in convenience and cost.
I guess there will be programmers needed for legacy systems in Fortran an Java. But new companies 5 years ago didn't use Fortran or Java for new projects, the coming companies will not hire coders for their products.
[0] https://www.economist.com/special-report/1999/10/07/cutting-...
> but what if AI creates much better, more powerful, exciting, better films than those you could imagine?
Well, something to deal with when it can. Right now it cannot do this.