@d7af9dea AI models are dependent on labeling(humans), retraining(expensive) active learning techniques (highly complex).  

This is all time intensive and expensive in an enterprise where you need to make sure this stuff never fails in production.

If you can sell it, then no problem... but that IS the problem currently!  There is potential here, but the economics are challenging.  Developers don't understand this ( @efa2e20e ). Dudes Named Ben and system architects do, all too painfully. 

 @d711be3b 

you have outdated knowledge that doesn't apply to generative AI

in other words:  you are clueless about generative AI