Four Teams, One Problem

Four runs, same room, paid to compare notes for an hour.

The pattern scales past the page. Four teams attacking one problem from four different angles, paid to compare notes for an hour at the end, would compound a problem-space faster than four companies running one experiment each and waiting for the knowledge to leak between them through attrition. People will turn up for that — expenses paid, an hour together, an award and a check. They are already running the experiments on their own time, in second jobs and hobby builds, and most of that work is wasted because nobody puts the four runs in the same room.