[kkfx]

Do not take me wrong, I appreciate and thanks anyone who contribute to FLOSS, but all low/no code approaches I see turn out to be garbage. IMVHO the reality is that people need to be trained and became capable of fishing alone instead of giving them fishes all days.

ML in ETL is needed for raw initial classification of documents received in various formats from various sources, to clean-up scanned crap, no more than that, all the effort to plug LLMs was so far and i bet will be for the next 10 years a disaster.

ETL is something that should not exists in a modern world because we should exchange data in usable formats instead of having to import the with all sort of gimmick, we do not have such acculturated world but at least we can try to simplify and teaching instead of adding entropy.