There is no inherrent reason why no new programming languages should be developed in the age of AI. As long as people are programming, there will be new languages. As there are still people playing traditional musical instruments in the age of sampling, there will still be programming humans, even if a lot will be automated. A programming language is a suitable way to communicate intentions and solutions with an LLM. There will be new languages which improve on this. But there will also be new "traditional" programming languages. I'm just developing one since more than a year because I see a specific need. Whether it will prevail and you can get a job using it is yet another, unrelated question. As in the past it will be mostly luck whether it will be accepted by a wider audience.
I was thinking about this the other way- new languages will always be designed by humans or human/ai collaboration but the task of a certain kind of viability will involve producing sufficient useful training material- and this when humans are intended to be able to also read the code strikes as being a new kind of gate, as such material will have to be automatically produced which may involve arriving at algorithmic definitions of fuzzy human syntactic concepts that refer to taste and readability. Super interesting design problem.
There is no inherrent reason why no new programming languages should be developed in the age of AI. As long as people are programming, there will be new languages. As there are still people playing traditional musical instruments in the age of sampling, there will still be programming humans, even if a lot will be automated. A programming language is a suitable way to communicate intentions and solutions with an LLM. There will be new languages which improve on this. But there will also be new "traditional" programming languages. I'm just developing one since more than a year because I see a specific need. Whether it will prevail and you can get a job using it is yet another, unrelated question. As in the past it will be mostly luck whether it will be accepted by a wider audience.
I was thinking about this the other way- new languages will always be designed by humans or human/ai collaboration but the task of a certain kind of viability will involve producing sufficient useful training material- and this when humans are intended to be able to also read the code strikes as being a new kind of gate, as such material will have to be automatically produced which may involve arriving at algorithmic definitions of fuzzy human syntactic concepts that refer to taste and readability. Super interesting design problem.
[dead]