You guys have to fix this , some weeks ago there was an issue in gemini where the model simply gave up when no more solutions were found to solve an issue , and Devs instructed the model to not give up on anything , well this is an issue , because not giving up specially on complex programming forces the model to make mistake after mistake and then loop all over again .
To instruct model to not give up you have to give it a lot of more training data , specially on programming . i am building a project with gemini pro and we are on the last bugs , on this image [Video player]i told gemini the issue and told it to when player starts up to automatically resize the video to the window app , over 4 different sandboxes i have tested the AI goes around and around trying to solve it and guessing what may be the issue but it never gets it fixed , and then loops all over again over the same issues it tried before with minor changes in it . The issue here is that AI really needs more training data because it is unable to find a proper solution to the issue , this image could give an impression that it is a simple project , however this app not only uses apis to server as also displays subtitles , it is a complex code project witch leads the AI on trying to fix a particular issue to make a new one ahead . I am doing this with gemini pro , with flask version i would never even try it , that flask model is only good to chat about things in life and have an hard personality . Please give it more books about programming to AI .