[jrm4]

Absolutely. Which is why what I think I'm saying is that -- it's not ncessarily important who or what WROTE the paper, it's simply important that human-in-the-loop is strong.

[jsrozner]

Right, but even careful reading cannot validate the experimental setup.