Take a second to go upvote You Are A Brain if you haven't already...
Back? OK.
Liron's post reminded me of something that I meant to say a while ago. In the course of giving literally hundreds of job interviews to extremely high-powered technical undergraduates over the last five years, one thing has become painfully clear to me: even very smart and accomplished and mathy people know nothing about rationality.
For instance, reasoning by expected utility, which you probably consider too basic to mention, is something they absolutely fall flat on. Ask them why they choose as they do in simple gambles involving risk, and they stutter and mutter and fail. Even the Econ majors. Even--perhaps especially--the Putnam winners.
Of those who have learned about heuristics and biases, a nontrivial minority have gotten confused to the point that they offer Kahneman and Tversky's research as justifying their exhibition of a bias!
So foundational explanatory work like Liron's is really pivotal. As I've touched on before, I think there's a huge amount to be done in organizing this material and making it approachable for people that don't have the basics. Who's going to write the Intuitive Explanation of Utility Theory?
Meanwhile, I need to brush up on my Python and find a way to upvote Liron more than once. If only...
Update: Tweaked language per suggestion, added Kahneman and Tversky link.
Of those who have learned about heuristics and biases, a nontrivial minority have gotten confused to the point that they offer Kahnemann and Tversky's research as justification for the self-arbitrages they've set up!
Suggested alternative wording:
"Of those who have learned about heuristics and biases, a nontrivial minority are so confused as to point to the biases research as justifying their exhibition of a bias!"
Completely agree that people just use methods such as tabu search a*, etc... without understanding them at all, same happens with machine learning techniques, or even statistic ones. Mostly they get by using the recommended algorithm/meta heuristic for the domain they are working at.
I strongly recommend python for doing this, it is the best language to begin programming with, I have several programs I did by myself, I can collaborate with the project.
1JGWeissman
I strongly agree. As an anecdotal data point, I understood the suggested alternative but not the original wording. And it is a powerful point to miss because I haven't heard of Kahnemann and Tversky.
Also, if mentioning specific researchers were central to the point, I would recommend linking to a resource about them, or better yet, create entries for them on the Less Wrong Wiki and link to those.
Take a second to go upvote You Are A Brain if you haven't already...
Back? OK.
Liron's post reminded me of something that I meant to say a while ago. In the course of giving literally hundreds of job interviews to extremely high-powered technical undergraduates over the last five years, one thing has become painfully clear to me: even very smart and accomplished and mathy people know nothing about rationality.
For instance, reasoning by expected utility, which you probably consider too basic to mention, is something they absolutely fall flat on. Ask them why they choose as they do in simple gambles involving risk, and they stutter and mutter and fail. Even the Econ majors. Even--perhaps especially--the Putnam winners.
Of those who have learned about heuristics and biases, a nontrivial minority have gotten confused to the point that they offer Kahneman and Tversky's research as justifying their exhibition of a bias!
So foundational explanatory work like Liron's is really pivotal. As I've touched on before, I think there's a huge amount to be done in organizing this material and making it approachable for people that don't have the basics. Who's going to write the Intuitive Explanation of Utility Theory?
Meanwhile, I need to brush up on my Python and find a way to upvote Liron more than once. If only...
Update: Tweaked language per suggestion, added Kahneman and Tversky link.