Probably Reddit.

Specifically https://www.reddit.com/r/shortcuts/ since it appears to be the last one standing.

You probably would have to make an account with a valid personal email, just to trace your prompts.

I wish I could downvote this multiple times.

ong bruh

There’s also virustotal, any.run, probably a few others outside of GitHub/gitlab scans

A vibe coded thought experiment, interesting but likely not worth the tokens to duplicate work.

I just skimmed the fully vibe coded page and saw the quote “Machines now speak every human language better than most humans.”. Interesting, and maybe possible if languages never change.

Hmm, nice page, I may end up vibe coding my own showcase page

Your loop is wrong, as it will go ‘intent -> llm(?) -> result’. Which doesn’t need to compile, but the llm its own black box of magic function that may or may not reproduce the same result based on intent. Also I would think the number of tokens would be concerning, as your example prompt looks 3x larger then the code snippet.

Are you worried about someone training their own model and recreating their own bot off of your public logs?

May be a great compliment if someone does try, haha. But seriously, the real value isn't in the text since it is very specific to my firm and work experience, the value is the underlying design logic and the specific 'if-then' professional constrains baked into the system. Plus, the audit logs are part of a 'transparency' experiment, if it helps someone else build a better tool for the industry, Im all for it!

It’s great to hear that! Hopefully you get the compliments you deserve after the logic is potentially reversed.

*What makes you think it hasn’t already been pulled/used in training from GitHub?

Just re-train your own model on your own collected data. No problem there.