Comments by "Nick" (@NicodemusT) on "Who's Solving The DEVELOPER SHORTAGE Crisis?" video.

Talking about testing software while early AI is writing ever more complex software seems a lot like re-arranging the chairs on the Titanic. Eventually, this sort of task will be a legacy art form more than a transferrable skill. I don't think AI is going to be writing tests. I understand everyone's apprehension about AI "taking our jobs", and I do think there is a reason to have that apprehension – but I am skeptical that the future that includes AI, even as a helpful tool, would encourage any overhead – while a LLM is pre-tuned to do that work ahead of time, and deliver pre-tested code before it's even written to file.
2
@vitalyl1327 today. Not in the future. Desperately hanging on to something you've convinced yourself is immensely meaningful is actually depressing. Why do you think LLMs need testing? Do you think that extra work is for computers or for humans? I'll give you a hint: it's not for computers. Altman has said that AI will eventually skip language altogether and talk directly to the processor. Where do you see pandantic testing fitting in here? Why would a self-learning system need testing in the long term? I hope your response includes acknowledgement of the next decade and not just what you perceive tomorrow.
1