It's not about being "Smarter", one of the main issues is "Consistency" and "Drift".
The same lack of consistency or Drift, that exists in visual/generative AI for image/video models, also exists in coding models.
Where the model will sometimes Hallucinate code methods/props for certain code frameworks/languages that don't even exist.
When they concretely solve the "Drift/Consistency" issue, then and only then will it be something that seasoned developers will have to worry about
"Replacing Them".
Right now most of the "Generative Code" AI platforms can create "Apps", but once it is time to "Refactor" or create additions to the app, unless you have some sort of "safe guards" in-place, there always stands the chance that the coding model, removes/renames functions/methods/props breaking previous working code, and even the model with change UI (if that is part of the code being worked on), removing key UI elements.
There is also the issue of unnecessary complexity, that some coding models introduce to solve simple coding problems.
you think CEO's and bean counters are receiving your message?
.
... I could sit back all day and not do a thing, obviously i have to know the codebase, but with the right prompts
