Some loose thoughts:
Use the word "illegal" more often. Most countries in the world, including the US, take a very dim view of marijuana. I think the Dutch drug policy is better, but getting caught may have serious consequences which should be thoroughly considered.
A high-quality mediator ("confessor") and illegal drugs (which implies quite a bit of preparation) are nontrivial resources. You also have to instruct the group. Wouldn't it be more fair to compare the effectiveness of your scheme to, say, a group with a high-powered brainstorm consultant (in addition to an "unaided" group)?
Are high-quality mediators useful in unaided groups? They are not common, at least. If they make the group less effective, is the extra creativity worth enough to overcome the bad effects?
I thought LSD was used in this way more often than marijuana?
More generally, are legal(-ish) substances able to deliver the same effect, if any?
Are you proposing that marijuana will help rationalist groups, or all groups? In the first case, define "rationalist".
Full disclosure: I'm Dutch, and I think mind-altering chemicals are squicky.
P.S. I don't understand your reddit link. Your last two links are broken. Check the rest.
EDITed to add: note that this article previously appeared in discussion, which I didn't know. Consider reading fiddlemath's comment and other comments.)
First allow me to show you my preliminary data.: http://www.reddit.com/r/bestof/comments/obtqe/its_they_live_these_are_reddit_robots_programmed/
Introduction
I think I may have found a novel use for an old technique, which may or may not have implications for rational decision making. I am open to constructive criticism or even deconstructive criticism if you make a sound argument. Ultimately, I would like the experiment to be put to the test. If you have the supplies and know-how to carry it out, then feel free to try it and report your findings.
The Goal:
Methods:
Some evidence that the Cannabis route might be a good one to pursue (more references to be added):
Anecdotal evidence:
Conclusion
References
What I'm missing. To be included later: