Models that have reached their limit should appear in a faded color, and when selected, the subscription can be upgraded

Contribute Feature
,
1

As you know, when we impose a limit on a model, a user can still pose questions to it even after exhausting the limit, only to be met with a message indicating that their limit has been reached. This is demoralizing and a waste of time.

image
image1310×632 59.4 KB

My suggestion is this: the models can continue to be listed, but if the user’s limit has been reached, that particular model could appear in a faded color. When the user selects it, a modal window could automatically prompt them to upgrade their subscription. This is essentially how the usage model should function. Someone who has reached their limit could increase it while asking their question, or even before, and then their question could be processed. This would be quick and convenient.

Furthermore, if we could also implement this, I believe it would be even better: the user’s remaining limits should be visible next to the model names. (Wow, this would be incredibly useful!)

What are your thoughts on this matter?

4 Likes
2

My general thoughts are that we want to slowly move towards removing this decision matrix from most users - in an ideal world they should be able to just talk to a robot and the engine underneath should figure out what the best llm is to serve the request.

It does add some overhead but also eliminates some decision paralysis non advanced users have.

What does “upgrade subscription” mean in this context, upgrade subscription to Discourse instance so you get higher limits, nudge admin to reconfigure llm?

There are a fair amount of assumptions built in here.

3 Likes