Tracking progress per problem

- Good way to mitigate against chatgpt. Problem solutions that are very often or always completed on first try are very suspicious
- The positive way of framing this would be to simply track the number of problems solved on the first try. This number can be recorded clearly in the profile and submitted as part of the self-grading
- A more sophisticated implementation would analyze finer grain details. Can we track time using the application? Or time between submissions? Any insights to be had there (and communicated to the user to help guide their practice?).