Hooked up the comic generator to #Gemini 1.0; I don't have access to 1.5 yet. Integration was easy. Working with the REST api is very similar to OpenAI. Only text generation was used, image generation is still done with #DallE3 here. As for the result, the model seems to write much more conservatively. Dialog lacks the cuteness #GPT4 provides. Oddly enough, the model seems to be very self deprecating too.
