Scalability! Observability! Evaluation!
The three most important things I can think of right now!
While five years back I used to lose my shi* while trying out any new fancy ML concept worrying that how would I write code for it.
@tapasadhikary I always do.
Whenever I find myself making up things while answering a question I say upright โI am not sure, havenโt got a chance to work on itโ or โI am not sure I am just thinking out loud hereโ generally.
@svpino One can easily build something but can they fix it just as easily when it doesnโt work?
When one doesnโt have the understanding what would they tell the AI to fix?