Comment by throwaway1851

Bard hasn’t been using Google’s best language models. I believe it just got an upgrade, however, and I’m now getting output that is significantly more coherent and useful than ChatGPT’s. It’s also a helluva lot faster, though that could owe to the limited access.