I'd call it an early experiment and what we could afford at the time to evaluate candidates for a SWE summer internship. Maybe other companies will like this approach and offer challenges that pay much better, with different setups & time horizons. Ex a user turnt his challenge into a 3-month contract-to-hire arrangement that paid $15k to the person who did the work & got the job.

Ultimately it's about getting the right signals. Real work produces the best signals. The rest of the mechanics are a WIP :)

Really appreciate your input, thanks a lot!!