Was This Man Responsible For Inventing The Computer?

00:00:00.000 | So the topic I want to discuss might seem unusual at first, but there is a backstory here.

00:00:05.520 | I will, I will quickly tell you here it is.

00:00:07.440 | Did Alan Turing invent the computer?

00:00:11.920 | All right.

00:00:13.400 | Why are we talking about this?

00:00:14.480 | Well, it goes back to last week.

00:00:17.960 | I was a guest on Sam Harris's podcast making sense.

00:00:23.480 | I don't know if you've heard that one yet, Jesse.

00:00:25.320 | Uh, but it's an episode that's creating quite a stir.

00:00:29.720 | Not because of me, but because the episode is, uh, begins with a 45 minute

00:00:36.360 | monologue from Sam, where he explains why he decided a few days before

00:00:41.920 | that episode posted to quit Twitter.

00:00:44.080 | So here is where, if we had a, uh, applause sound effect, a confetti and

00:00:49.960 | applause sound effect is where, where we would, where we would hear it.

00:00:52.880 | So Sam quit, uh, Twitter and it sort of made sense to attach his

00:00:59.080 | announcement about that to this interview.

00:01:00.680 | I'd recorded with him a few weeks earlier, because obviously when you

00:01:03.000 | talk to Cal Newport, you're going to, you're going to hear a lot about

00:01:05.360 | tech and tech and society and not necessarily a very positive view of Twitter.

00:01:09.240 | Now, interestingly in that interview, later in the interview, and we're

00:01:14.360 | talking about Twitter, Sam is cataloging a lot of his concerns about it.

00:01:19.120 | And I took a swing and made a pitch in the interview directly.

00:01:24.120 | Sam, you should quit Twitter.

00:01:25.520 | Now I don't get to take credit for that.

00:01:28.280 | Sam actually specifically addresses this at the end of his monologue.

00:01:31.600 | He says, as you will hear, I already had doubts when I was doing this

00:01:35.360 | interview and Cal was pushing me to quit it.

00:01:37.120 | I didn't quit Twitter because Cal told me to, but he was one of the voices

00:01:41.120 | in my head when I made the decision.

00:01:42.720 | So I'm going to take, let's take a partial W yeah, on that Jesse.

00:01:47.080 | Um, but anyways, obviously I'm glad he did it, not just for his own sanity,

00:01:50.680 | but I think because it's a role model.

00:01:52.160 | Uh, as long time listeners.

00:01:56.680 | No, my issue is not with the existence of these social media platforms.

00:02:01.520 | It is with the assumption of ubiquity.

00:02:03.320 | That I think is what the problem is.

00:02:04.960 | The idea that everyone has to use the platform.

00:02:07.760 | That's what I, that's what I might.

00:02:09.440 | I don't care that Twitter exists.

00:02:10.800 | I care that it's a big deal that I don't use it.

00:02:13.120 | And so when we have more high profile people like Sam opt out, it opens

00:02:18.720 | up that possibility to others, others who maybe feel like the.

00:02:22.160 | Cost or outweighing the benefits.

00:02:24.200 | It makes it easier and easier for those who follow in his wake to say, you know

00:02:27.400 | what, I'm going to do something similar.

00:02:29.320 | I got a lot of heat when I originally left Twitter, Sam, who's way more

00:02:33.280 | well-known to me is getting a huge amount of heat right now, but more of us

00:02:37.000 | who go through this, the easier it will become for those who follow.

00:02:39.400 | So I think that's all good news.

00:02:40.640 | Anyways, it's a long interview.

00:02:44.080 | A lot of it, we're dealing with tech issues, a lot of tech and

00:02:47.000 | society, a lot of tech criticism.

00:02:48.240 | Very interesting stuff.

00:02:49.040 | A lot of new theories you haven't heard me talk about before.

00:02:50.840 | Worth listening to.

00:02:52.720 | But earlier in the show, Sam and I wandered across a bunch of esoteric

00:02:57.720 | topics that just sort of popped to our head as we were chatting.

00:03:00.520 | And one of the topics that came up relatively early in the conversation

00:03:05.200 | was this thought experiment that Sam had considered before, where he was

00:03:11.240 | thinking if I could go back in time, somewhere between the 1930s, the 1940s.

00:03:17.080 | And, uh, let's say kill a single individual.

00:03:21.000 | If my goal was to delay as long as possible, the development of the

00:03:25.040 | modern computer, which individual would you kill now, obviously that's a,

00:03:29.640 | maybe a, a, a violent construction for what's an actually the very interesting

00:03:33.400 | question, you know, who was probably most singly influentially responsible for

00:03:37.520 | the development of the modern computer.

00:03:39.040 | I thought that was a cool thought experiment.

00:03:40.600 | So we got into it.

00:03:41.360 | Now, one of the names that often comes to people's mind when they think about the

00:03:46.120 | invention of the computer is Alan Turing.

00:03:50.320 | And as Sam and I talked about, I think Turing gets too much credit as a, uh,

00:03:58.200 | initiator of the development of modern digital computing.

00:04:01.080 | I'm a huge Alan Turing fan.

00:04:02.640 | I teach Turing to our doctoral students at Georgetown.

00:04:05.680 | I know his work very well.

00:04:07.120 | He's the father of theoretical computer science and incredibly

00:04:10.040 | influential thinker and a very original thinker, but his role in the invention

00:04:17.440 | of the computer, I think has been inflated in recent decades, he would

00:04:20.880 | not be, in other words, my choice of who to go back and, and, uh, rub out.

00:04:24.920 | If I was trying to delay the development of the computer.

00:04:27.160 | Now I'll tell you soon who I think that person is, but first let's

00:04:31.360 | return to the question of Turing.

00:04:34.120 | So what did he do that became so connected to computing in the modern minds?

00:04:41.240 | Well, it really comes down to the notion of the Turing machine.

00:04:44.240 | If you want to understand the notion of the Turing machine, and I promise you,

00:04:47.080 | I'm not going to get into professor mode here.

00:04:48.440 | I'll be very brief, but if we want to get into the notion of the Turing machine,

00:04:51.960 | you have to go back to this paper he wrote called on computable numbers and

00:04:56.320 | their connection to the Einstein problem, which is a German name for a problem that

00:05:01.440 | was posed in the late 19th century by David Hilbert.

00:05:04.520 | Now this problem had nothing to do with computers.

00:05:07.200 | This is from the 1800s.

00:05:08.440 | But what it asked is, can we come up with what they would call back

00:05:14.240 | then an effective procedure?

00:05:16.800 | Today, we might call this an algorithm, but back then they would

00:05:19.680 | call it an effective procedure.

00:05:20.560 | That is a step-by-step series of instructions for solving any

00:05:24.960 | math problem we might want to solve.

00:05:26.400 | Does every math problem have a step-by-step way to solve it?

00:05:30.200 | This was a big question in mathematical logic.

00:05:32.320 | A lot of people were working on it and Turing came up with an answer.

00:05:37.360 | And the way he came up with an answer is he said, let's have a formal

00:05:40.880 | definition of an effective procedure.

00:05:43.000 | And that's when he came up with this thought experiment of the Turing machine.

00:05:46.160 | It's a set of instructions, an infinite tape, a read head that can move from

00:05:52.080 | position to position on the tape, read what's there, look up in the instructions

00:05:55.880 | what to do, maybe overwrite what's there, move one direction or the other.

00:05:59.600 | Turing made this argument that this abstract machine in theory could implement

00:06:04.600 | any possible effective procedure.

00:06:06.520 | So every effective procedure has a corresponding Turing machine.

00:06:11.880 | He then did a bit of mathematical logical tricks where he said, look, we could

00:06:15.160 | describe any such Turing machine with a sequence of whole numbers.

00:06:19.440 | And we could just put those whole numbers together and just get a really big whole number.

00:06:23.120 | So every Turing machine and therefore every effective procedure has

00:06:26.520 | a corresponding whole number.

00:06:28.520 | Now it might be really big.

00:06:29.640 | It might be a couple hundred thousand digits long, but just conceptually speaking,

00:06:33.720 | there is a way to label every possible effective procedure with the whole number.

00:06:38.480 | Then he looked at what do we mean by a problem?

00:06:41.960 | And he focused in on a subset of problems you might try to solve.

00:06:45.560 | These were called decision problems.

00:06:46.800 | He did a little bit of mathematical logic and he argued every problem can be

00:06:53.440 | represented by a real number.

00:06:55.480 | That is a number, a decimal point number that has an infinite number of decimal places.

00:06:59.800 | You know, 1.0146578 often to infinity.

00:07:04.840 | And in fact, there's a one-to-one correspondence there that you could, you

00:07:07.520 | could, you could take every real number and that exactly describes a particular decision problem.

00:07:13.480 | This was a big deal because there is a well-known result going back to Cantor.

00:07:20.640 | Now we're going back to the 19th century that says there are many more real numbers

00:07:24.040 | than natural numbers.

00:07:24.880 | There exists no way to map every natural number onto a real number such that you've

00:07:30.880 | covered all the real numbers.

00:07:34.400 | The impact of that is, okay, if we, if we map every possible effective procedure to

00:07:39.760 | the problem it solves, there'll be many problems left over that aren't being mapped

00:07:43.720 | to by an effective procedure.

00:07:46.040 | Math, math, math, logic, logic, logic.

00:07:48.960 | And the conclusion is there's many more problems out there in the universe and

00:07:52.520 | there are algorithms or effective procedures that can solve them.

00:07:54.840 | Most things that most problems out there can't be solved by effective procedures.

00:07:59.640 | This was the question that Hilbert was trying to answer.

00:08:02.480 | Turing answered it.

00:08:04.320 | So this was all about logic, mathematical logic, foundational math that was going

00:08:09.440 | on right then.

00:08:09.960 | None of this had to do with computers.

00:08:11.640 | The reason why we connect this to modern digital computing is you can say Turing's

00:08:18.960 | notion of a Turing machine is an abstract notion of a computer because you have this,

00:08:25.440 | that the tape could have on it instructions that a Turing machine could run.

00:08:29.480 | He talked about in his original paper, something called the universal Turing

00:08:32.560 | machine, where the input on the tape is a description of another Turing machine and

00:08:36.120 | it simulates it.

00:08:36.760 | So you do have some of the conceptual basics there of a computer reading a

00:08:39.640 | program and executing it.

00:08:41.320 | Okay, fair enough.

00:08:42.800 | Also, we do know that von Neumann at Princeton was familiar with Turing's work.

00:08:47.760 | He met Turing when Turing was visiting the Institute for Advanced Study in

00:08:51.640 | Princeton.

00:08:52.280 | Von Neumann later advanced the von Neumann architecture for modern computers, which

00:08:56.400 | is the one we use today.

00:08:57.360 | So there's a little bit of an influence there as well.

00:09:00.640 | But the idea that Turing single handedly sort of introduced this idea that we could

00:09:05.120 | have these universal computing machines, that's just not true.

00:09:07.680 | Before Turing even did this work, well before this work was well known outside of

00:09:13.520 | esoteric mathematical circles, we already had general purpose analog electronic

00:09:18.840 | computers.

00:09:19.360 | We had, for example, Vannevar Bush's differential analyzer at MIT.

00:09:22.920 | In the mid 1930s, we got the very first, we began to get the very first ideas being

00:09:28.720 | proposed for making fully electronic computing machines.

00:09:32.320 | As the war went on, there was a huge push to have more advanced electronic

00:09:36.760 | computing.

00:09:37.160 | They were using these to calculate artillery tables and to help aim at the

00:09:41.680 | aircraft guns and some sort of cybernetic sensors.

00:09:43.720 | A huge research effort for this.

00:09:45.440 | And while it's true that Turing after the war got involved in a project in the UK to

00:09:51.920 | develop an electronic computer, this was one of at least a half dozen ongoing

00:09:55.560 | projects, many of which finished sooner.

00:09:58.760 | I think the ENIAC at Penn, for example, there was a Van Nuyman's project at

00:10:02.760 | Princeton.

00:10:03.240 | There was a project going on at Harvard.

00:10:05.040 | A lot of people were working on this problem.

00:10:07.600 | And they didn't need Turing to do it.

00:10:10.600 | The final thing people point to is they saw that that movie about whatever it was

00:10:17.240 | called, the imitation game.

00:10:19.240 | And like, well, didn't he invent these sort of computing machines to break the

00:10:23.240 | enigma code?

00:10:24.280 | No, those were developed by the Poles.

00:10:26.760 | The Polish code breakers developed those.

00:10:29.200 | Turing was just building a more advanced version of those machines.

00:10:32.000 | They had more funding.

00:10:33.280 | So they used the initial work that the Poles had put into breaking the enigma

00:10:37.600 | and then they expanded it.

00:10:38.360 | I love Turing, but he didn't invent the computer.

00:10:42.080 | So who would I go back and rub out if I was trying to delay the computer?

00:10:47.960 | I would say Claude Shannon in the early 1930s.

00:10:54.000 | Claude Shannon in the early 1930s wrote the most important master thesis that

00:10:59.600 | anyone has ever written before.

00:11:01.120 | It was called a symbolic analysis of relay and switching circuits.

00:11:04.240 | This master thesis is what figured out the entire field of digital electronics.

00:11:09.320 | This was the really key breakthrough that everything else was built on.

00:11:14.360 | Shannon had been interning at Bell labs where he was seeing electromagnetic

00:11:21.000 | relays, phone networks use electromagnetic relays to automatically connect calls

00:11:25.840 | using electrical signals.

00:11:27.080 | He was also studying for a degree in mathematics at MIT.

00:11:31.280 | He put those two things together and he said, wait a second, you can take purely

00:11:37.440 | logical statements expressed in Boolean algebra and you can implement them with

00:11:41.360 | electronic circuits using these electromechanic relays.

00:11:45.040 | So you can take an arbitrary mathematical specification of a logical circuit and

00:11:49.640 | build it, anything you can come up with, any Boolean algebra statement you can

00:11:52.720 | come up with, we have a systematic way of building that with wires and magnets.

00:11:59.640 | We can build an electronic circuit.

00:12:01.400 | I have a quote.

00:12:02.880 | He said this later in life, it just happened that no one else was familiar

00:12:06.800 | with both of these fields at the same time.

00:12:08.720 | So he happened to be in both worlds, math, phone company came together.

00:12:13.880 | That was probably the single biggest innovation because once we realized we

00:12:18.200 | can build arbitrary logic into electrical circuits, that's what opened up the whole

00:12:22.640 | hope that whatever idea we have that we want to implement, whatever adding circuit

00:12:26.760 | or logic circuit or whatever we need to implement our conceptual design of a

00:12:31.160 | computer, whatever we can come up with, if we can specify it mathematically, we can

00:12:34.600 | build it.

00:12:35.000 | So if we're going to follow this sort of oddly Marshall exploration of early

00:12:42.240 | computing, Shannon in the thirties, getting rid of Shannon in the thirties would

00:12:46.320 | probably have a bigger impact.

00:12:48.080 | Then getting rid of Turing in the thirties.

00:12:51.240 | [inaudible]

00:12:57.240 | [inaudible]

00:13:00.240 | (upbeat music)

Was This Man Responsible For Inventing The Computer?

Chapters