As Vaniver said:

Konkvistador brought up the idea about two months ago, but to the best of my knowledge the idea is still being researched.

If you are interested in doing more reading on the topic of MOOC or rationality education related literature or perhaps want to contribute something to the exploration of the subject PM me an email adress and I'll add you to the list of people who can edit & read the common notes our informal research group is accruing on google docs.

I'm a very long-time lurker (OB days) who's moved to post for the first time, since this is slap bang in my professional area of expertise. (My username is my real name, which readily searchable.)

The idea of a MOOC on rationality is a really interesting one.

I think it has a potentially very large payoff, but a low chance of success.

What's the potential benefit?

It's an obvious opportunity. You could reach an audience of tens of thousands or more to learn your methods, and an order of magnitude or two more who'd be vaguely aware. Technology enthusiasts and people who are open to new approaches are particularly attracted to MOOCs. So the potential audience for a MOOC is - I think - precisely your target audience. You could reach a scale at a speed you can't possibly do otherwise.

The everything-goes-right scenario is that you get an awful lot of people to engage with your project, who become long-term adherents to your cause, and go on to attract even more people. This really would "raise the sanity waterline" in a way and at a scale that would make a substantial difference.

What's the risk?

It's hard to do right, and it's public. You have no experience of doing it (and indeed have very little experience of teaching, relative to other MOOC sponsors) so will almost certainly make "obvious" mistakes. That could end up making you look bad and incompetent, leaving you a bit worse off than if you hadn't tried.

In theory, you could improve faster than other providers, if you were better able to apply a rational, empirical approach. The pedagogy of most MOOCs is rubbish (technical term of art there), and there is great scope for doing it better. I expect success will come faster to operators with baseline expertise rather than those bootstrapping from next to nothing, but I could be wrong - those who already "know" about teaching may be burdened with mental obstacles to doing it much better. However, there are a lot of extremely smart people working in extremely effective feedback loops in this space, including serial entrepreneurs, so I think your relative advantage here is not likely to be significant.

I don't think the risk of getting it wrong is a huge problem, though: it's an experimental space, so failures are expected and tolerated, and it's unlikely that you'd get widely slated in a way that really hurt your reputation unless you did something really daft. I think a much more likely failure mode is a lack of interest, leading to a lack of numbers.

Your offering is likely to be unattractive. It's a crowded space. And there are two big factors working against you.

First, reputation. You don't really have one, or at least, not one for excellence in teaching. You are going head to head with Harvard, MIT, Stanford et al. Taking a MOOC is unconventional (for now), so having the reassurance of a conventional, blue-chip university is a big attraction that you are in no position to supply.

Second, your subject area looks to me highly likely to prove unattractive. You are very convinced that your methods and techniques can dramatically improve people's success in life, but hardly anyone else is convinced (yet?). A proven, high-status course with obvious, direct and widely-understood relevance to a high-paid career is going to look much more appealing to an audience who aren't already convinced of your principles. And there are a lot of existing MOOCs that meet those criteria, with more coming.

You could try to mitigate this factor by focusing specifically on topics with widely-accepted applicability - probability and risk seem an obvious area. But this brings out the reputation issue very starkly: your claim to being experts in the area is not prima facie a strong one compared to established educational institutions, and your approach will likely seem idiosyncratic to those with some existing conventional knowledge.

What are the costs?

The main direct cost is the time to understand what's involved, prepare the material, and support the course. You don't need me to tell you to consider what you could otherwise spend that time on.

There's the technology itself. This isn't very hard cutting-edge stuff, but will require some technical resource, which again has an opportunity cost.

There's also the risk of losing potential big-ticket paying customers for your face to face sessions. (The self-cannibalisation issue.) That's a possibility, but not a big danger. I think a successful MOOC is likely to prove more of a shop window for the expensive, high-input product than a competitor to it. I would expect a successful CFAR MOOC to increase the number of takers for bespoke CFAR face to face workshops and courses. And it'd let you reach a much wider audience than you possibly could with face-to-face sessions. Certainly this appears to be the thinking of the likes of Harvard, MIT and Stanford.

So is it a good idea?

In summary, I think it's unlikely to succeed, but if it does, the payoff is potentially vast: basically, you'd achieve the lofty goals you've set yourselves.

How attractive a proposition that is depends on how you like to evaluate that sort of offer.

(Bias disclaimer: I'm inclined by disposition, training and experience to poke holes and find flaws in arguments and proposals for educational technology projects.)