Note: The following outline of my research proposal is unfinished. I posted it in the discussion section to spur conversation and get constructive criticism (successfully, I might add). If you have any suggestions, then please make them. I will be monitoring the discussion and improving the proposal until I feel it is ready to be posted as a main article.


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:

  •  Catalyze the brainstorming process in a way that increases both the number and quality of ideas made.


  • Find a problem that needs solving. Unifying general relativity and quantum mechanics is a good, but ambitious, example. Some more likely problems that could be solved are: "How might I solve my relationship problems," or "how can I advertise my company's product to its target demographic", or "what are some ideas to make quick money."
  • Find 2-3 rationalists who understand the problem well. They don't need to be expert rationalists; the most important part is that they know the difference between rationalization and rationality. In the QM example above, and in most scientific applications of the method, all players should have access to the experimental data.
  • Assign two of the three rationalists to the "brainstormers" group (name subject to change), whose primary concern is to make logical connections between the data to form hypotheses.
  • Assign the odd-rationalist-out as the Confessor, whose primary concern, like in TWC, is to preserve sanity. It is the Confessor's job to catch the brainstormers when they make a logical leap or use biased reasoning. Some tactics the Confessor might use are the rationalist taboo, the reversal test, and argument from the least convenient world.
  • This is a scaled up version of what the brain seems to do. We need the brainstormers and the Confessor to act the part of the Apologist and the Revolutionary, respectively.
  • The Confessor - brainstormer dynamic is interesting in its own right, but I believe it can be improved. Now bear with me, because the optional step is for the brainstormers to smoke Cannabis. Not too much, but just enough so that connections between ideas are more quickly apparent to them. Remember, the goal is to have the brainstormers make many connections. They need to output quantity over quality, while the Confessor picks out anything that is quality and gently guides the brainstormers toward more quality ideas. Think of it like r-selected evolution.
  • Ideally, we would split twelve rationalists into four groups of three (the alternative is to use the same group repeatedly). Group 1 would be told to just brainstorm the problem. Group 2 would be told to choose one among them to be the Confessor. Group 3 and 4 would be told the same, but their brainstormers would be given either Cannabis or a placebo.
  • A placebo can be made by extracting the cannabinoids using ethanol or glycerine. All that should be left after extraction is plant matter, and the tincture can be used later for medicinal or recreational purposes. The placebo Cannabis and active Cannabis will have to be rolled into joints because extraction removes some of the plant's pigmentation. If you have access to a lab, then you might follow  this procedure  for the extraction; otherwise, use  this guide  for doing the extraction at home.
  • If you don't want to go to the trouble of making the placebo, then you may skip the control group and only do groups 1 - 3. It would be nice to get some preliminary data, even if skewed slightly by the placebo effect.
  • For data collection, the Confessor will note down any idea made by the brainstormers, marking the ones which were discarded. After a given amount of time, enumerate the data and compare the groups. The hypothesized result is that the smoking group will make the greatest quantity of ideas, followed by the non-smoking partitioned group, followed by the normal brainstorming group. It is also hypothesized that the smoking group will make the greatest quality ideas, due to a combination of the highly creative nature of ideas made while high (explained below) and the Confessor's job of immediately discrediting any faulty reasoning.

Some evidence that the Cannabis route might be a good one to pursue (more references to be added):

  • There is evidence that Cannabis engages the mind in semantic Hyper-Priming1,2, meaning that distantly-related concepts are primed quickly after having been exposed to an idea. For instance, a smoker might quickly respond to the word "fish" with "submarine," whereas someone who is sober might respond with "fin." If I understand it correctly, then this doesn't mean the smoker cannot answer "fin," only that the more distantly related concepts are given a higher priority than they normally would. One can see why this might be advantageous for brainstorming, but I suggest to take my - and's - interpretation of the paper with a grain of salt until someone with access can read it in full.
  • *Cannabis allows erroneous perspectives to be rapidly dismissed in the light of new evidence. While high, it is easy to put one's pride aside and say "oops". This is made especially easy if the user has cached understanding of the art of rationality. In other words: they will listen to the Confessor.* <-- (I haven't found any literature to support this claim, yet. It seems true in my experience, but it might not be true for everyone. If the brainstormers prove to be too clingy, we could alter the method by changing the Confessor's name to Kiritsugu and having the brainstormers agree to always defer to the Kiritsugu's better judgement. The Kiritsugu will have to take care to examine its own judgement and only discard the truly irrational ideas).

Anecdotal evidence:

  • Artists, writers, and even scientists have long used Cannabis and other psychoactive drugs as a tool to make "insights." I'm defining insight as the connection and/or creation of ideas (erroneous or otherwise), possibly due to hyper-priming. The Confessor, in the early pioneers' case, was usually their sober self. As Hemingway wrote, "write drunk; edit sober."
  • Less gifted stoners have been doing this for ages  but they - for the most part - are completely undisciplined, believe in dubious pseudoscience, and/or don't have a rational observer to moderate them.
  • This is going a bit meta, but the outline to the outline of this idea was made while I was high. It was the first time I smoked since having been introduced to Less Wrong and "The Way", and I was surprised to find that I still had most of my wits about me. Although I would often begin down paths that were just Rationalizations, I usually caught myself. In the instances where I didn't catch myself, and it seemed like a legitimately good insight, I wrote the idea down for future (sober) consideration.
  • One of the good, practical, non-meta insights I made that night was a life plan. My plan up until this point had been to finish my undergraduate degree and then immediately go to grad school, relying on my schooling and a bit of luck to maybe hopefully turn into a somewhat-successful scientist somewhere along the road. The problem is that I suffer from quite a bit of procrastination, in part because I don't know exactly what I want to do. I don't have any strong passions or any real motivation. My college career, so far, has been an uphill battle against crippling akrasia.
  • Aided by Cannabis, I finally saw the obvious: I need to make an effort to find a passion. My new plan is to get a job as a computer programmer after finishing undergrad, but to continue self-teaching in Biology and other sciences. I've already taken the first step by having Computer Science as my minor, and I can help my resume along right now by getting involved in open source projects. As for self-teaching, that's made easy by open courseware like that found on Khan Academy, MIT, and other places, and I always have the old-fashioned solution of just reading textbooks. After following my interests for a while and learning what things I really, really like to learn about, then I'll go to grad school with an actual PhD thesis in mind and money in the bank.
  • I'm attributing these insights (the life plan, some other ideas I'm not mentioning, and even the hypothesis itself) to hyper-priming and later editing, but they might have just been made because I was focused on the problems. Hence the need for an experimentally-controlled test.


  • Cannabis allows connections to be made between concepts which normally seem unrelated. This is an experience commonly reported by users, and experimentally verified. Some of these connections will inevitably be false, but others might be true, and a third party - a Confessor - might be able to distinguish truth from falsehood. Whether the Confessor - brainstormer dynamic is any more efficient or productive than a normal brainstorming session is an open question, and the only way to really know is to test the hypothesis.


How cannabis makes thoughts tumble. (n.d.).Mind Hacks. Retrieved from
Morgan, C. J. A., Rothwell, E., Atkinson, H., Mason, O., & Curran, H. V. (2010). Hyper-priming in cannabis users: a naturalistic study of the effects of cannabis on semantic memory function. Psychiatry Research, 176(2-3), 213-218. doi:10.1016/j.psychres.2008.09.002

What I'm missing. To be included later:

  • References to the benefits and techniques of traditional brainstorming. In lieu of that, for now, here's  this  and  this .
  • More references to Cannabis research.


New Comment
17 comments, sorted by Click to highlight new comments since: Today at 11:04 PM

I'm not the only one getting a somewhat cult-like vibe from this, am I?

But really, this seems incredibly impractical. A few Lesswrongians meeting up and getting high are not going to solve the world's problems. Not that Cannabis might not be of some assistance to actual experts, but simply being rational does not immediately qualify you to solve very much.


I'm not the only one getting a somewhat cult-like vibe from this, am I?

I'm only using concepts created by members of the community to help ensure understanding of the material. I'm in no way attached to the idea of "x-rationality" and "rationality dojos". This method could be easily utilized by anyone with a cursory understanding of cognitive biases.

A few Lesswrongians meeting up and getting high are not going to solve the world's problems

Again, maybe not, but if the tool is shown to be even 5% more effective than a normal brainstorming process, then won't it be worth it?

Not that Cannabis might not be of some assistance to actual experts, but simply being rational does not immediately qualify you to solve very much.

I really overrepresented the kind of skill you need to have to make use of Cannabis as a tool for decision making. You don't need to be Eliezer Yudkowsky, lukeprog, James Randi, or even me (and I'm, admittedly, extremely irrational and unknowledgeable. I'm working on it). You only need to know the difference between a rationalization and rationality. All the other stuff is really, really useful, don't get me wrong, but I think that's the minimum to use the technique. And the technique, again, is not just to get high... it is to have a rational, sober observer moderating the conversation.

Bah, now you're making me want to rewrite the whole thing without Eliezer's style of cultish countercultishness. It's damned ironic, since when I first discovered Less Wrong I was a bit worried of the same exact thing. Thanks for telling me about the elephant in the room.

EDIT: It reads a lot more sane, now, I think. I've sanitized it of most of the superfluous LessWrong terminology. Remember, I came up with the idea using the less-efficient solo method of "write [high]; edit sober." If I had a Confessor with me to keep the sanity in the first place, I expect it would have read a lot less "cultish." I still think the underlying idea is a good one, and worthy of experiment.

Overall, I think this sounds like a great idea to test. in fact, the idea of having group discussions include a formal confessor, who focuses just on keeping the argument clean, is an excellent idea on its own, even without psychoactive substances.

That said, I suspect I have a decent idea of why you're getting downvotes. The rest of this are some of my suspected reasons why. Thus, this will probably read like biting criticism. I'm hoping you take it in the constructive spirit it's intended.

Find 2-3 rationalists who know every prior needed to solve the problem.

I suspect you mean to find some rationalists who strongly have the background knowledge that's likely to be necessary. An important aspect of being confused about a problem is not actually knowing what pieces are useful to think about at the same time. "All the necessary priors" is something you learn after you've solved the problem.

Assign the odd-rationalist-out (preferably the most rational of the three)

In any group of several "rationalists" that I know, it's not all that easy to pick the "most rational" person. There are many different skills involved in x-rationality, and the ability to spot likely confusions and likely biases - to referee a discussion, essentially - is an important but not all-consuming subskill. On the other hand, I really like the idea of having a designated "rationalist referee" for serious discussions. However, this:

It is the Confessor's job to determine whether the hypotheses created by the brainstormers have high Bayesian probability, and that all logical leaps and/or biases are identified as soon as they are made.

is probably impossible. At best, it's the Confessor's job to watch for missing logical steps or biased arguments, and point them out as soon as noticed. We have to make the Confessor out of a person, after all.

I'd add that the Confessor should suggest clarifying tactics (eg., rationalist taboo, employing the reversal test, or arguing from the least convenient world) where they might be useful.

It's not a huge deal, but by "prior" we usually mean the probability one ascribes to ideas before weighing evidence, rather than the ideas themselves. Your use of "prior" here is probably part of what's set off people's rejections.

Moreover, if you look closely at it, much of your "evidence that this will work" isn't much evidence; it's a shiny anecdote broken into bullet points. The parts that are evidence, though, are only evidence for "cannabis aids unfiltered insight," rather than "this technique will work." (I have my own reasons to suspect that the rest of this technique will work, but I've been putting off the long, well-researched post about brainstorming that I ought to be able to link here. :/ )

The conclusion, as stated, seems vastly overconfident. "Teamwork" is already part of the idea of a rationality dojo. This leaves "cannabis", then, as your stated "missing ingredient" to making a "rationality dojo" work. This... is likely to be a vast overstatement -- not least because it claims only to help the whole group reach better conclusions, rather than spurring the group to act on those conclusions, or getting everyone in the group to internalize those conclusions. Better to say, frankly, that this could help, and that you'd like to test this. (I'm pretty sure that's what you mean! But it's not how it reads.)


Thank you. That is exactly what I meant to say, even if I didn't know I meant to say it ;) I'm making the changes you suggested.

EDIT: Done

[This comment is no longer endorsed by its author]Reply

Why stop with cannabis? If you want to max out creativity go for LSD.


Maybe. I don't have any experience with LSD, so I really can't say. The goal is to max out creativity while still being able to listen to your inner (and the outer, the Confessor) rationalist. If you can drop acid and still do that, then it might be an even better tool for days when you can afford to trip a while.


May I just remark that Cannabis is dangerous; It is known to occasionally induce psychosis on first time use and that susceptibility to said phenomenon is highly unpredictable.

A cannabis psychosis is usually permanent, and as far as I can deduce, very detrimental to rational thinking. I will list sources and relevant studies if prompted to do so.

Indulge with resposibility :)

Do you have a source for this? I did a quick search, but couldn't find anything obvious to my untrained eye. Some links suggest that there's a link, but that the causality might run in both directions.


Here is a recent German study strongly linking onset of Schizophrenia and other psychotic symptoms with cannabis use:

And an older Swedish one (I can't seem to find the full article):

Mainly it seems there is a correlation for those with pre-existing mental problems, or those who are disposed to mental illness but have no symptoms. I advise again that it is unpredictable who are in the latter group.

ETA: Wikipedia has an okay "state of the science" article:

There is usually a couple votes worth of variance, especially when a post first comes up. I live a bit too far from other LWers to implement this myself, but upvoted for taskifying very well.

The first improvement I would like to see is some references for the claims about cannabis. They're common enough, and research on such things is notoriously hard to come by, which is why I still upvoted without, but it would make the point much stronger.

The next suggestion I'd make is that your "solving quantum physics problems" example could probably be replaced by something that general LWers could do.

Otherwise this seems like a good discussion post and I'd love to see someone else try this and report back on it.

upvoted for taskifying very well

Hmm, disagree. This book is an adequate taskification of results-focused meetings. Compared to that, the above post leaves me with an impression of 1) brainstorm, 2) ??? 3) Profit! A large part of the post is about the non-conventional advice to use THC, and that could use more scholarship: the THC-creativity link has been investigated, and I would expect a LW post making substantive claims about it to point to some relevant papers.

The above post leaves me with the impression of a particular structure for group brainstorming. I say it is taskified very well because I could easily (i.e. within 2 hours probably) implement the procedure and notice the difference and attempt to measure the results compared with brainstorming without following the procedure.

I agree that it needs more scholarship; but I expect different things out of discussion posts than I do out of published books.


Yes, sorry, I'm fleshing it out now. This is the discussion area for unfinished ideas, isn't it? I'm taking everyone's criticisms into consideration.

Sounds like a decent hypothesis, go try it. Want a five-dollar grant or something?


Haha. I've already tried it with a couple of non-xRationalists with great success (read the life plan above); the problem is that a) I spent most of my brainstorming time explaining to them how and why they were making cognitive biases, and b) I don't know any rationalists in real life. I'm currently converting my older brother and his girlfriend.

We'll also need to do a double-blind control, with the brainstormers smoking something that seems like Cannabis but isn't psychoactive. Is there something like that? Maybe we could use ABV. A second control could also be useful, where the brainstormers don't smoke anything.

And, of course, we'll want as much data as possible. That means multiple groups conducting the experiment and reporting their findings. Like you said, it's inexpensive, so it shouldn't be too hard to get the experiment rolling so long as enough people are interested.

EDIT: The following is sort of irrelevant now, since people have begun giving me constructive criticism. I'm keeping it for posterity's sake.

EDIT: Please explain your downvotes. I was led to believe that this community was less susceptible to the follies of the representativeness heuristic (I'm assuming that is a large reason for the downvotes, given MixedNuts' first comment). Does it sound like the meta-belief of Cannabis held by uneducated, delusional hippies that the plant will cure all of the world's ills, and that is why you dismiss it? If you're not downvoting it just because it sounds like another pipe-dream made by a crazy stoner, then ask yourself why you're downvoting it. Examine why the reasoning is bad and then tell me why, please.

I'm quoting the Less Wrong introduction thread

However, it can feel really irritating to get downvoted, especially if one doesn't know why. It happens to all of us sometimes, and it's perfectly acceptable to ask for an explanation. (Sometimes it's the unwritten LW etiquette; we have different norms than other forums.) Take note when you're downvoted a lot on one topic, as it often means that several members of the community think you're missing an important point or making a mistake in reasoning— not just that they disagree with you! If you've any questions about karma or voting, please feel free to ask here.

It is very likely that it is not you, but me who is missing something. I would like to be elucidated.

EDIT2: I got my explanation, see above. I'm glad this will evolve into a discussion.

Intending this as gently as possible, I'm wondering, WingedFoe, if you see anything ironic about this bit in your text and wider stereotypes about the long term results of pot use on human character...

The problem is that I suffer from quite a bit of procrastination, in part because I don't know exactly what I want to do. I don't have any strong passions or any real motivation. My college career, so far, has been an uphill battle against crippling akrasia... Aided by Cannabis, I finally saw the obvious: I need to make an effort to find a passion.

More references to Cannabis research.

Hard to come by because of the legal restrictions. The best sources I have seen:

Altered States of Consciousness edited by Charles Tart, 1969, Wiley.

Pharmako/Poeia by Dale Pendell, 1995, Mercury House.

They include pros and cons although it is obvious both guys are at least a little more pro than con. From Pendell's book: "Smoking it occasionally makes you wise; smoking it a lot turns you into a donkey." (p.199)

New to LessWrong?