I wonder how much a search on Stackoverflow would have improved the algorithm. To me, AI for coding is like Stackoverflow with a much faster feedback loop, but without the community comments that often give very useful context.
odd, I find the feedback loop is slower than just searching, given how slow AI is. But then again I never manage to get prompt correct in the first couple attempts.
In any case AI is much faster than asking a question on Stackoverflow and waiting for a reply. I agree that finding answers to existing questions on Stackoverflow takes about the same time as asking AI.
but being able to get EXACTLY UNT EXACTLY what you want... youre gonna need to be able to code, for that.
Also just pasting in code from an AI... it might work, but as soon as you try to make a larger system that communicates with other systems, youre gonna have a bad time.
and lets not forget.. do these non-coders know how to make sure youre code is safe, and all that jazz
Wondering whether this development will decrease the high level of arrogance that one encounters across developer teams (not always of course, but the tendency is evident). Yes, programming is much more than coding and not everything can be built or fixed by AI, at least yet. It's a sure thing we will get there, though. Would be phenomenal to not hear "we must first write a user story, do some grooming, evaluate the tools, then do more grooming, and talk even more before we start" in the future.
I think the mistake is in assuming that coding is coding. The fact is AI is introducing a whole new segment to the coding arena by encouraging enthusiasts and hobbyists to develop their random ideas. Yes the code is undoubtedly horrible, but the results kind of work. And again, this is the worst it's going to be.
These micro apps aren't going to be massively life-changing, at least not at first, but as the AI develops so too will the results. The big deal about this is suddenly application development is not locked into a cycle controlled by 'experts'.
Similar to the way the PC unlocked the data processing department stranglehold. Remember the days when you had to go to your DP manager to get a report printed on the centralized printer?
We're going to see a tsunami of apps arriving from across the world, all sorts of scratch-my-itch solutions, in various qualities and languages. For example - https://showmeyourbolt.io/projects
Whether you think that's a good thing or a bad thing probably depends on your 'religion'.
9 comments
[ 2.5 ms ] story [ 30.4 ms ] threadbut being able to get EXACTLY UNT EXACTLY what you want... youre gonna need to be able to code, for that.
Also just pasting in code from an AI... it might work, but as soon as you try to make a larger system that communicates with other systems, youre gonna have a bad time.
and lets not forget.. do these non-coders know how to make sure youre code is safe, and all that jazz
These micro apps aren't going to be massively life-changing, at least not at first, but as the AI develops so too will the results. The big deal about this is suddenly application development is not locked into a cycle controlled by 'experts'.
Similar to the way the PC unlocked the data processing department stranglehold. Remember the days when you had to go to your DP manager to get a report printed on the centralized printer?
We're going to see a tsunami of apps arriving from across the world, all sorts of scratch-my-itch solutions, in various qualities and languages. For example - https://showmeyourbolt.io/projects
Whether you think that's a good thing or a bad thing probably depends on your 'religion'.