I'm a founder. I think rationality training has helped me in a few major ways:
I have internalized that "nothing works or fails by magic" -- I expect that there is a mechanism of action for pretty much any process, and I just have to understand it, and I've started building a toolkit for generating these understandings/models.
I have started to become more strategic. I still fail at this in many ways, but these days, I often think to: look at the bigger picture; look for alternate routes to achieve my goals; consciously consider whether my current goals are worth achieving, or whether they have become lost purposes.
Communication: Asking for examples from my cofounders has become an ingrained, extremely common habit when I'm not sure I understand them. I also often address the meta-level of communication, often asking "why are we having this discussion?" or observing "this meeting isn't going anywhere".
Not trusting myself: I set timers, I put things on schedules, I write things down, I block myself from websites. I know the ways in which my brain isn't as awesome as I wish it were, and so I take effective steps to patch these weaknesses.
(I'm going to go write a blog post about this now, because it seems valuable enough to keep.)
Part 1 is up. http://techhouse.org/~lincoln/blosxom.cgi/rationality/reflections.html
I ran out of time for now, but I intend to post part 2 soon. I'm glad you replied to me, because I probably wouldn't've taken the time to write this otherwise. Hopefully you'll reply again to remind me to post part 2 :)
Why fight what you can use?
Given that biases are an inherent part of the human nature, you might consider discussing how to use them to the startup's advantage. This does not have to be Dark Arts, though if you want to succeed in business, almost anything goes.
In fact, if entrepreneurs understood probabilities, most would never start any business, given the prevalent rate of small business failure. Furthermore, startup employees tend to be underpaid and overworked, hoping for a large payoff down the road, which almost never happens. You can get more out of people by building a sense of loyalty (a totally misplaced concept for an employee).
On the darker side, one can use anchoring to pay people as little as a frugal but eventually successful startup ("we have to be like them to succeed!"), In fact, if you go through the list of cognitive biases on wikipedia, most of them can be used to the startup's founders' advantage.
Of course, it is essential to not unwittingly succumb to the known biases, if you are running a business (sunk cost is the most pervasive one). One standard approach is to pay a knowledgeable outsider to question and critique your decisions on a regular basis. This used to be one of the function of the board of directors, but we all know how well this works out in practice (those who stir the pot do not keep this cushy job for long). The VCs are rarely up to the task, either.
Success of the startup depends a lot on the industry it is in:
http://www.amazon.com/Illusions-Entrepreneurship-Costly-Entrepreneurs-Investors/dp/0300113315
That book is actually pretty optimistic for Silicon Valley type stuff if you read it.
Startups that get venture capital tend to have high expected values:
http://80000hours.org/blog/12-salary-or-startup-how-do-gooders-can-gain-more-from-risky-careers
To be frank, I think most entrepreneurs are naïve and incompetent (this is from hanging out in online entrepreneurship communities and going to events for entrepreneurs). As an entrepreneurial type, most entrepreneurial types annoy me. (The stereotypical example would be a college entrepreneurship club that doesn't result in the creation of any new companies. Also, try any nonprogrammer that wants to start a web company, or anyone who complains about being rejected from Y Combinator – being rejected isn't a sure sign of failure, but complaining probably is – the appropriate response is surely either stoicism or defiance :D) I'm sure that a driven less wrong user has a significantly better chance of succeeding than the median entrepreneur.
This might be a good start http://web.archive.org/web/20090211154634/http://blog.pmarca.com/2008/03/the-psychology.html
Using the outside view can be really powerful, the trouble is that it's difficult to not just go "well, it will be different this time." I don't know of any well-tested way to communicate the outside view in a way that people actually use it.
Many of the blog posts on this blog are good:
The big thing about startups is that it's very easy to be either too optimistic or too pessimistic about how likely your idea is to succeed, since you're so emotionally invested and since there is so much genuine uncertainty. To combat this, watch for when your estimate of your success prospects change based on limited or no evidence (for example, may be an uninformed friend seems pessimistic) and remind yourself to work based on the moving average of your estimates of how successful you will be (average together how optimistic you were on each of the past three Mondays when deciding how hard to work, assuming you haven't acquired any genuine new and important information in that time).
I think a good general algorithm for startups might go something like this: Write out a description of how you're going to make money, figure out what the biggest uncertainty or pain point is, and then figure out the quickest and cheapest way to learn more about that uncertainty or pain point. Then adapt your plan based on what you've learned and move on to the second biggest uncertainty or pain point.
In case you didn't see this Paul Graham post:
http://lesswrong.com/lw/9ik/a_word_to_the_resourceful/
More stuff:
http://lesswrong.com/lw/2p5/humans_are_not_automatically_strategic/
http://lesswrong.com/lw/fc/you_are_a_brain/
I'd suggest emphasizing how frequently people are overconfident, and browsing less wrong by tag to find more relevant posts. I see rationality as more of a set of attitudes than a set of skills (I don't know that much, some of my thought processes may be defective, etc.)
One aspect of rationality is acting optimally (reasonably) on given data. In startups, you often need to act on little data, because when you wait for perfect information, the opportunity is gone. Unknowns can still be handled rationally, but need to be treated probabilistically. From this point of view, I really like the Lean Startup method, because it provides a framework to addressing the unknowns: lay out your plan, identify the greatest unknown, test it efficiently, rinse and repeat.
Re: biases and startups:
On the flipside, biases can also be taken advantage of, for example, when selling to people, using anchoring, framing, etc.
The inside vs. outside view seems to me to be one of the most useful tools available for people running a company.
Delusions of Success is a good article about the inside and outside views, written specifically for business types.
I have been given the opportunity to prepare a workshop for the General Assembly team in London. General Assembly is geared towards education of entrepreneurs and aspiring entrepreneurs and have been very successful in New York, now expanding to London. The workshops are 90 minutes long, and usually gather anywhere from 15 to 35 people who have paid to attend.
While I considered doing something on concrete coding skills, I think by far the superior alternative (for myself and the audience) is to do a crash course on cognitive bias as it relates to startups, maybe throw in some other topics on rationality in a similar context. I am fairly confident that startups are an excellent testing ground for extreme rationality as they require exceptionally quick assimilation of new skills and knowledge, as well as demand rapid decisions with incomplete information.
So, as part of the brainstorming for this, here are my questions for you:
1.Do you think educating startup founders on cognitive bias/rationality will help them improve their outcomes?
2.Which biases would especially affect startups? Which of these can be mitigated (either by knowing about them or by utilising explicit strategies)?
3.What is a good way to use 90 minutes to get this information across?
4.What prior material exists to introduce rationality in a fast-paced manner? What prior material exists that relates startups to rationality?
5.Other relevant thoughts welcome
Should I go ahead with this, I will of course make the deck available for any others who may want to do similar presentations elsewhere.