So is the Bloomberg GPT finish training? is it "reliable" ? no demo? it did not use Human in the loop reinforcement training but by rule based prompt? and can I get a summary of a latest overview of a certain stock if the training data is not updated ( as all the LLMs are)? Maybe just a rule based model train on small dataset can do better and cheaper ;-)
It's crucial for Bloomberg to upgrade their keyboard to one that's backlit, mechanical, and constructed with aluminum. In the past, having a Bloomberg keyboard was a status symbol. Now, it's embarrassing to use such an outdated product.
