This:
Deeku is an AI persona, and o1-mini is a model by OpenAI. But why is that model visible and changeable? Is the purpose of that so I can change the model no matter what I’ve configured on personas? And when it is that way, will it mean that Deeku will use o1-mini, even if it should use gpt4o? Because when I change a persona, that one doesn’t change.
Models are changeable there mostly as a way for developers to test personas against multiple models easily.
Most production sites will only have a single model configured, which will render that moot. And it is in our roadmap locking personas to a single model by default soon.
So this is a temporary situation then? Because I’m using different personas to do, of course, very different tasks, but also to limit use. And the last thing I want to happen is a regular user trying to chat with o1 when there should be gpt-4o mini. That actually happened to me when I was totally lost when a persona, that should use o1, failed badly. It took a while and several attempts before I realized it was using gpt-4o mini instead.
