You would have to be a lot more clear if you want real help with this.
- No AI has learning abilities currently.
- They only function by “mimicing learned information” because the entire conversation context window is being passed with each message.
- If you aren’t getting the results your after, it’s simply because you aren’t using the system well for the way that it’s designed.
- If your getting better results elsewhere, and you want advice on how to actually get specifically different results using GPT4o or other models available on chatgpt.com, you would have to be precise about:
- What prompts are you issuing and testing across models/providers?
- What are the difference in results are you getting and what do you like/not like about the various results?
- Are you open to changing your prompting method to try and achieve different results across different models?
Each provider provides a different model that is going to produce different responses. However, none of them provide a “learning ability” nor a “memory” function. This is all a completely “hidden aspect” of how the middleware system from the provider actually works. It’s not a fundamental aspect of the “AI or LLM model itself”.
