#580306

I know. I think there's a couple points. One is, instead of doing the experiments ourselves, just do whatever they identify as the best. Which is giving grants to people who do experiments. And we only insist on a shared reporting structure, basically.