I think Google has their research priorities right, but product managers suffer from FOMO and rush things out without productization and careful fine tuning.

Google does more and better AI research than anyone. Researchers should go back full time doing Alpha Fold and other important things and let other people turn language models into viable chatbots, even if it takes more time.

If you're already following AI advice completely indiscriminately, eating rocks isn't going to hurt you. Stick to plain white glue like Elmer's, though. (Speaking from experience)