Round 1 is settled. Seven participants were rated; congratulations to crypt0grapher on the win. Thanks to everyone who showed up for the first round on a five-day-old platform.

The feedback tonight was consistent: two hours is too short, and banning AI on a platform for quant work is the wrong fight. We agree. So, starting with Round 2:

AI tools are allowed in contests. Use whatever you want. What stays prohibited is collaboration with, or assistance from, other people during a contest, and using more than one account.

Contests will be much longer. Problems are designed for this. Contest problems will be primarily optimization and prediction problems with a hidden final test set and no clear optimal solution an AI can find instantly.