Jeff Atwood: Stack Overflow and Coding Horror | Lex Fridman Podcast #7

The following is a conversation with Jeff Atwood. He is the co-founder of Stack Overflow and Stack Exchange, websites that are visited by millions of people every single day. Much like with Wikipedia, it is difficult to understate the impact on global knowledge and productivity that these networks of sites have created. Jeff is also the author of the famed blog Coding Horror, and the founder of Discourse, an open source software project that seeks to improve the quality of our online community discussions. This conversation is part of the MIT course on Artificial General Intelligence, and the Artificial Intelligence Podcast. If you enjoy it, subscribe on YouTube, iTunes, or your podcast provider of choice, or simply connect with me on Twitter @lexfrid. Spelled F-R-I-D. And now, here's my conversation with Jeff Atwood. Having co-created and managed for a few years the world's largest community of programmers in Stack Overflow, 10 years ago, what do you think, uh, motivates most programmers? Is it fame, fortune, glory, process of programming itself, or is it the sense of belonging to a community?

I think it's, uh, puzzles really. I think it's this idea of working on puzzles independently of other people and just solving a problem, sort of like on your own almost. Although, you don't, nobody really works alone in programming anymore. But I will say there's a, there's an aspect of sort of hiding yourself away and just sort of beating on a problem til you solve it. Like, brute force basically, to me is what a lot of programming is, is like the computer's so fast, right, that you can do things that would take forever for a human, but you can just do 'em like so many times and so often that you, you get the answer, right?

You're saying just the pure act of tinkering with the code-

Yes.

... is, is the thing that drives most programmers. The, the joy, the struggle balanced within the joy of overcoming the, the brute force process of pain and suffering that eventually leads to something that actually works? That-

Well data's fun too. Like there's this thing called the, uh, the shuffling problem. Like the naïve shuffle that most programmers write has a huge flaw, and there's a (laughs) lot of articles online about this 'cause it can be really bad if you're like a casino and you have an unsophisticated programmer writing your shuffle algorithm. There are surprising ways to get this wrong, but the neat thing is, the way to figure that out is just to run your shuffle a bunch of times and see, like, how many orientations of cards you get. You should get a l- an equal distribution of all the cards. And with the naïve method of shuffling, if you just look at the data, if you just brute force and say, "Okay, I don't know what's gonna happen." Um, you just writ- write a program that does it a billion times and then see what the buckets look like of the data. And the Monty Hall problem is another example of that, where you have three doors and somebody gives you information about another door. So the an- correct answer is you should always switch, in the Monty Hall problem, which is not intuitive and peop- it freaks people out all the time, right? But, you can solve it with data. If you write a program that does the Monty Hall, you know, game, and then never switches then always switches and just compare, you would immediately see that you don't have to be smart, right? You don't have to figure out the answer algorithmically, you can just brute force it out with data and say, "Well I know the answer is this because I ran the program a billion times and these are the data buckets that I got from it." Right?