There are several commenters in this thread who actually see it as having zero software engineering team at all; there would not be a single person on staff who could read and understand application code.
I would agree about needing fewer heads performing types of roles, and I could even buy that tech staff would hardly ever need to handwrite code directly.
For serious projects where critical data, physical safety, etc for end users are at stake, I still don’t see the path toward simply having no in-house engineer to certify changes generated by an LLM
I would agree about needing fewer heads performing types of roles, and I could even buy that tech staff would hardly ever need to handwrite code directly.
For serious projects where critical data, physical safety, etc for end users are at stake, I still don’t see the path toward simply having no in-house engineer to certify changes generated by an LLM