v1.04.76

Taking a break from psychology notes to do some coding again.

• began working on a curriculum tool so educators can easily build em without having to work through the complicated json structure. its a bit hard since i want to export a zip, and im not entirely sure how that would work, and also i need to double check how my course.js processes the json so i can get exactly all the information to put into it.

• ai detector thing. examines github repositories and find all comments and flags exessive amounts, or exessive explanations in them (classic ai fingerprints) as well as flagging verbose function names and shows the time it would take to make the project at 70 wpm and detects inconsistencies in commit times (eg 1000 new code typed after 2 minutes of another commit that would be flagged) then gives a % of confidence. 0-30% is prob human, 30-50% is mixed unsure, 50-100% is definetly full ai. Before you accuse someone, make sure to do througouh like questioning and interviews and such because this tool is imperfect. This way i can easily identify projects that have disgenuine work and report them to stardance to take care of. (https://cassetu.github.io/Novara/html/aisubTool.html) I got gemini to help reformat the code since it was a bit of a mess, dw i had a full disclaimer. (ai vibe coders take notes.) It was a big challenge getting the github token to work, but it functions now and doesn’t steal your information!! :D

• also i ran my own project through it once, it was clear, and then i ran it again after commiting it, and now it thinks my project is ai, because the code scans for these specific traces, that are present in the code itself, meaning that if it scans itself, it will find everything its looking for. T - T (false positive in the first image)

leave a comment to help! thx!