SCAMP Tutor

The ethos of SCAMP is one of owning and understanding the tools of your craft. AI coding pulls in two different directions in this respect. In terms of ownership, most of the frontier AI models are closed-source and closed-weight, meaning that you, as a user, cannot own them, only rent them. On the other hand, they are becoming highly capable of building bespoke tools and interfaces, offering the potential for you to build and maintain your own custom tooling.

From the standpoint of understanding, AI also pulls in both directions. The pernicious aspect is that it stands waiting to offer us a shortcut to any problem-solving task, happy to let us offload our critical thinking. When you vibe code, you leave all the details to the AI and never truly understand what’s going on under the hood. However, at the same time, if you approach AI coding assistants with a strong motivation to understand, they can be effective in answering all manner of questions, helping you to understand details rather than obscuring them.

In this spirit, I offer a kind of homemade AI SCAMP tutor: simply copy and paste the following message into your preferred open- or closed-source chatbot:

Read https://marcthespark.github.io/scamp-tutor/tutor.txt. I want you to be my SCAMP tutor,
using the instructions and resources described there. On a scale from 1-10, my experience
levels are as follows: coding 5, python 4, music theory 7, scamp libraries 2.

(Replace the four numbers with your own experience levels, from 1 = none to 10 = expert.)

The tutor has been instructed to guide you, at a level consistent with your current ability and understanding, to create SCAMP scripts that you understand. The file linked to is plain text — feel free to read it — and tries to encode principles of good teaching that encourage ownership and understanding.

The instructions also link to resources: an indexed, plain-text version of all of the documentation and examples on this website. This way, instead of hallucinating SCAMP-like code from its general training data, it is more likely to be helpful and knowledgeable about SCAMP specifically. If you are a more advanced coder and less in need of step-by-step tutoring, you can instead simply point your chatbot at the resources directly.

Happy SCAMPing, and please feel free to offer feedback on how this experimental tutor performs, both in terms of knowledge and its approach to teaching.