Donald Knuth: Alan Turing was the First 100% Geek | AI Podcast Clips
00:00:00.000 | - So you refer to people, including yourself,
00:00:05.000 | who gravitate toward a kind of computational thinking
00:00:09.220 | as geeks, at least I've heard you use that terminology.
00:00:13.620 | - It's true that I think there's something
00:00:16.340 | that happened to me as I was growing up
00:00:18.260 | that made my brain structure in a certain way
00:00:21.780 | that resonates with computers.
00:00:24.220 | - So there's this space of people,
00:00:26.060 | 2% of the population, you empirically estimate.
00:00:29.900 | - That's been-- - Proven?
00:00:32.860 | - Fairly constant over most of my career,
00:00:34.940 | however, it might be different now
00:00:38.220 | because kids have different experiences when they're young.
00:00:40.820 | - So what does the world look like to a geek?
00:00:45.540 | What is this aspect of thinking that is unique to--
00:00:50.540 | - That makes, yeah. - That makes a geek?
00:00:54.580 | - This is a hugely important question.
00:00:59.380 | In the '50s, IBM noticed that there were geeks
00:01:04.380 | and non-geeks and so they tried to hire geeks
00:01:09.780 | and they put out ads for the paper saying,
00:01:12.380 | if you play chess, come to Madison Avenue
00:01:14.660 | for an interview or something like this.
00:01:16.420 | They were trying for some things.
00:01:18.100 | So what is it that I find easy
00:01:22.060 | and other people tend to find harder?
00:01:24.580 | And I think there's two main things.
00:01:27.660 | One is this ability to jump levels of abstraction.
00:01:32.660 | So you see something in the large
00:01:38.580 | and you see something in the small
00:01:41.460 | and you pass between those unconsciously.
00:01:45.860 | So you know that in order to solve some big problem,
00:01:50.420 | what you need to do is add one to a certain register
00:01:54.820 | and that gets you to another step.
00:01:57.140 | And below the, yeah, I don't go down to the electron level
00:02:01.380 | but I knew what those milliseconds were,
00:02:04.220 | what the drum was like on the 650.
00:02:06.060 | I knew how I was gonna factor a number
00:02:09.700 | or find a root of an equation or something
00:02:13.100 | because of what was doing.
00:02:14.660 | And as I'm debugging, I'm going through,
00:02:17.140 | did I make a key punch error?
00:02:19.900 | Did I write the wrong instruction?
00:02:23.140 | Do I have the wrong thing in a register?
00:02:25.860 | And each level is different
00:02:30.540 | and this idea of being able to see something
00:02:34.780 | at lots of levels and fluently go between them
00:02:40.020 | seems to me to be much more pronounced
00:02:42.940 | in the people that resonate with computers like I do.
00:02:47.420 | So in my books, I also don't stick just to the high level
00:02:52.500 | but I mix low level stuff with high level
00:02:57.500 | and this means that some people think
00:03:01.380 | that I should write better books
00:03:06.820 | and it's probably true but other people say,
00:03:10.780 | well, but that's, if you think like that,
00:03:13.900 | then that's the way to train yourself.
00:03:15.860 | Keep mixing the levels and learn more and more
00:03:18.900 | how to jump between.
00:03:20.460 | So that's the one thing.
00:03:21.540 | The other thing is that it's more of a talent
00:03:25.300 | to be able to deal with non-uniformity
00:03:31.900 | where there's case one, case two, case three
00:03:34.260 | instead of having one or two rules that govern everything.
00:03:39.300 | So it doesn't bother me if I need,
00:03:44.860 | like an algorithm has 10 steps to it.
00:03:48.620 | Each step does something else that doesn't bother me
00:03:51.220 | but a lot of pure mathematics is based on one or two rules
00:03:55.540 | which are universal and so this means that people like me
00:04:00.540 | sometimes work with systems
00:04:03.260 | that are more complicated than necessary
00:04:04.900 | because it doesn't bother us
00:04:06.460 | that we didn't figure out the simple rule.
00:04:09.820 | - And you mentioned that while Jacobi, Boole, Abel
00:04:15.700 | and all the mathematicians in the 19th century
00:04:17.740 | may have had symptoms of geek.
00:04:21.700 | The first 100% legit geek was Turing, Alan Turing.
00:04:26.100 | - I think he had, yeah, a lot more of this quality
00:04:31.100 | than any, just from reading the kind of stuff he did.
00:04:36.540 | - So how does Turing, what influence has Turing had on you?
00:04:41.980 | In your way of thinking?
00:04:44.540 | - So I didn't know that aspect of him
00:04:48.220 | until after I graduated some years.
00:04:50.700 | As an undergraduate, we had a class
00:04:53.940 | that talked about computability theory and Turing machines
00:04:56.860 | and it was all, it sounded like a very specific
00:05:01.860 | kind of purely theoretical approach to stuff.
00:05:06.540 | So when, how old was I when I learned that he had
00:05:13.380 | designed machines and that he wrote the,
00:05:16.340 | he wrote a wonderful manual for Manchester machines
00:05:22.580 | and he invented all subroutines
00:05:27.580 | and he was a real hacker that he got his hands dirty.
00:05:32.580 | I thought for many years that he had only done
00:05:38.540 | purely formal work.
00:05:39.700 | As I started reading his own publications,
00:05:43.020 | I could feel this kinship and of course,
00:05:48.020 | he had a lot of peculiarities.
00:05:51.100 | Like he wrote numbers backwards because,
00:05:54.180 | I mean, left to right instead of right to left
00:05:57.580 | because that's, it was easier for computers
00:06:00.440 | to process them that way.
00:06:01.920 | - What do you mean left to right?
00:06:04.220 | - He would write pi as nine, five, one, four, .3, I mean.
00:06:11.940 | Okay. (Lex laughing)
00:06:15.340 | - Right, got it.
00:06:17.380 | - Four, one, .3.
00:06:18.700 | On the blackboard, I mean, when he,
00:06:21.500 | he had trained himself to do that
00:06:26.740 | because the computers he was working with
00:06:29.740 | worked that way inside.
00:06:31.060 | - Trained himself to think like a computer.
00:06:32.820 | Well, there you go, that's geek thinking.
00:06:34.980 | (Lex grunting)
00:06:37.560 | (upbeat music)
00:06:40.140 | (upbeat music)
00:06:42.720 | (upbeat music)
00:06:45.300 | (upbeat music)
00:06:47.880 | (upbeat music)
00:06:50.460 | (upbeat music)
00:06:53.040 | you