A repo to explore the dotnet intermediate exercises ( https://www.udemy.com/course/csharp-intermediate-classes-interfaces-and-oop/ ) by using a gpt LLM as the prompting solver.
The results of each exercise are posted at their respective .md files in the prompt_explore folder. As the whole of the exercise is solved using a gpt llm's response, the explanations will include the hurdles it faced and how prompting was used to improve the results. The initial prompts are the questions provided in the exercise, and are not discussed on further.
As a contextual model understands code prompts and answers to it, you can say it faces similar diffusion problems similar to the 5+ finger errors problems in image diffusions. Negative biases would easily improve a text-based code, but a noobie would face difficulty at finding the improvement areas crucial to the code. The understanding of fundamentals is required with the current state of Copilot GPT 4. Given this insight, it also explains the problems that might be faced when trying to create an app from scratch. Not knowing where the GPT model is going wrong could lead down an unnecessary debugging rabbit hole.
The exercise pdf files are available with the course and are not shared publicly for obvious reasons.