Guido van Rossum: Python
技术与编程AI 与机器学习生物与进化音乐与艺术历史与文明
📋 章节目录
暂无章节信息
🔑 关键词
pythonprogrammingdonlanguagegoingprogramintelligencecomputerrulesdesigncodeconsciousnesslanguagesthoughtableguidointerestingbuiltdidnsystems
💬 精彩语录
"be resolved while maintaining backwards compatibility or sort of using a very gradual path of evolution"
在保持向后兼容性的同时得到解决，或者使用非常渐进的进化路径
— Guido van Rossum （1:08:20.420）
"And of course the GIL connected to the parallelization, I suppose, the global interpreter lock problem."
当然，我认为 GIL 与并行化有关，是全局解释器锁问题。
— Guido van Rossum （1:19:59.040）
"or social networks and they say, well, algorithms, and that's like a totally different interpretation"
或社交网络，他们说，好吧，算法，这就像一个完全不同的解释
— Guido van Rossum （44:04.840）
"of every form of learning or sort of controlling computer systems should always be called programming."
任何形式的学习或控制计算机系统的行为都应该被称为编程。
— Guido van Rossum （48:30.320）
🎙️ 完整对话（867 条）
Lex Fridman (00:00.000)
The following is a conversation with Guido van Rossum, creator of Python, one of the most popular
以下是与 Python 的创建者 Guido van Rossum 的对话，Python 是最受欢迎的语言之一
Lex Fridman (00:05.680)
programming languages in the world, used in almost any application that involves computers
世界上几乎所有涉及计算机的应用程序都使用编程语言
Lex Fridman (00:11.120)
from web back end development to psychology, neuroscience, computer vision, robotics, deep
从网络后端开发到心理学、神经科学、计算机视觉、机器人学、深度学习
Lex Fridman (00:17.760)
learning, natural language processing, and almost any subfield of AI. This conversation is part of
学习、自然语言处理以及人工智能的几乎所有子领域。这段对话是
Lex Fridman (00:24.560)
MIT course on artificial general intelligence and the artificial intelligence podcast.
麻省理工学院关于通用人工智能和人工智能播客的课程。
Guido van Rossum (00:29.280)
If you enjoy it, subscribe on YouTube, iTunes, or your podcast provider of choice, or simply connect
如果您喜欢它，请在 YouTube、iTunes 或您选择的播客提供商上订阅，或者直接连接
Guido van Rossum (00:36.080)
with me on Twitter at Lex Friedman, spelled F R I D. And now, here's my conversation with Guido van
在 Twitter 上与我联系 Lex Friedman，拼写为 F R I D。现在，这是我与 Guido van 的对话
Guido van Rossum (00:44.720)
Rossum. You were born in the Netherlands in 1956. Your parents and the world around you was deeply
罗苏姆.你1956年出生在荷兰。你的父母和你周围的世界深深地影响着你。
Guido van Rossum (00:53.120)
deeply impacted by World War Two, as was my family from the Soviet Union. So with that context,
我和我来自苏联的家人都深受第二次世界大战的影响。因此，在这种背景下，
Lex Fridman (01:02.000)
what is your view of human nature? Are some humans inherently good,
你对人性的看法是什么？是不是有的人性本善，
Lex Fridman (01:07.360)
and some inherently evil? Or do we all have both good and evil within us?
还有一些本质上是邪恶的？或者我们内心都有善与恶吗？
Guido van Rossum (01:12.240)
Guido van Rossum Ouch, I did not expect such a deep one. I, I guess we all have good and evil
Guido van Rossum 哎呀，没想到这么深。我，我想我们都有善有恶
Lex Fridman (01:24.880)
potential in us. And a lot of it depends on circumstances and context.
我们身上的潜力。这在很大程度上取决于环境和背景。
Guido van Rossum (01:31.440)
Peter Bell out of that world, at least on the Soviet Union side in Europe, sort of out of
彼得·贝尔脱离了那个世界，至少在欧洲的苏联一侧，有点脱离了这个世界。
Guido van Rossum (01:38.800)
suffering, out of challenge, out of that kind of set of traumatic events, often emerges beautiful
出于挑战、出于一系列创伤性事件的痛苦，往往会显得很美丽
Guido van Rossum (01:46.480)
art, music, literature. In an interview I read or heard, you said you enjoyed Dutch literature
艺术、音乐、文学。在我读到或听到的一次采访中，你说你喜欢荷兰文学
Guido van Rossum (01:54.320)
when you were a child. Can you tell me about the books that had an influence on you in your
当你还是个孩子的时候。你能告诉我一些对你影响很大的书吗？
Guido van Rossum (01:59.760)
childhood? Guido van Rossum
童年？吉多·范罗苏姆
Guido van Rossum (02:01.520)
Well, with as a teenager, my favorite writer was my favorite Dutch author was a guy named Willem
嗯，十几岁的时候，我最喜欢的作家是我最喜欢的荷兰作家是一个名叫威廉的人
Guido van Rossum (02:09.120)
Frederik Hermans, who's writing, certainly his early novels were all about sort of
弗雷德里克·赫尔曼斯（Frederik Hermans）正在写作，当然他的早期小说都是关于
Guido van Rossum (02:19.440)
ambiguous things that happened during World War Two. I think he was a young adult during that time.
Lex Fridman (02:31.600)
And he wrote about it a lot, and very interesting, very good books, I thought, I think.
Lex Fridman (02:40.800)
Peter Bell In a nonfiction way?
Guido van Rossum (02:42.560)
Guido van Rossum No, it was all fiction, but it was
Lex Fridman (02:46.400)
very much set in the ambiguous world of resistance against the Germans,
Guido van Rossum (02:54.560)
where often you couldn't tell whether someone was truly in the resistance or really a spy for the
Guido van Rossum (03:03.840)
Germans. And some of the characters in his novels sort of crossed that line, and you never really
Guido van Rossum (03:11.280)
find out what exactly happened.
Lex Fridman (03:13.840)
Peter Bell And in his novels, there's always a
Lex Fridman (03:16.880)
good guy and a bad guy, the nature of good and evil. Is it clear there's a hero?
Lex Fridman (03:22.160)
Guido van Rossum No, his heroes are often more,
Guido van Rossum (03:25.120)
his main characters are often anti heroes. And so they're not very heroic. They're often,
Lex Fridman (03:36.640)
they fail at some level to accomplish their lofty goals.
Guido van Rossum (03:40.800)
Peter Bell And looking at the trajectory
Guido van Rossum (03:43.040)
through the rest of your life, has literature, Dutch or English or translation had an impact
Lex Fridman (03:50.560)
outside the technical world that you existed in?
Lex Fridman (03:54.160)
Guido van Rossum I still read novels.
Guido van Rossum (04:00.640)
I don't think that it impacts me that much directly.
Lex Fridman (04:05.200)
Peter Bell It doesn't impact your work.
Guido van Rossum (04:07.280)
Guido van Rossum It's a separate world.
Guido van Rossum (04:10.080)
My work is highly technical and sort of the world of art and literature doesn't really
Guido van Rossum (04:17.440)
directly have any bearing on it.
Lex Fridman (04:19.120)
Peter Bell You don't think there's a creative element
Guido van Rossum (04:22.400)
to the design? You know, some would say design of a language is art.
Lex Fridman (04:26.880)
Guido van Rossum I'm not disagreeing with that.
Guido van Rossum (04:32.160)
I'm just saying that sort of I don't feel direct influences from more traditional art
Lex Fridman (04:39.360)
on my own creativity.
Guido van Rossum (04:40.880)
Peter Bell Right. Of course, you don't feel doesn't mean
Lex Fridman (04:43.280)
it's not somehow deeply there in your subconscious.
Lex Fridman (04:46.000)
Guido van Rossum Who knows?
Lex Fridman (04:48.240)
Peter Bell Who knows? So let's go back to your early
Guido van Rossum (04:51.200)
teens. Your hobbies were building electronic circuits, building mechanical models.
Lex Fridman (04:57.440)
What if you can just put yourself back in the mind of that young Guido 12, 13, 14, was
Guido van Rossum (05:06.080)
that grounded in a desire to create a system? So to create something? Or was it more just
Lex Fridman (05:12.240)
tinkering? Just the joy of puzzle solving?
Guido van Rossum (05:14.720)
Guido van Rossum I think it was more the latter, actually.
Guido van Rossum (05:18.720)
I maybe towards the end of my high school period, I felt confident enough that that
Guido van Rossum (05:29.920)
I designed my own circuits that were sort of interesting somewhat. But a lot of that
Guido van Rossum (05:39.120)
time, I literally just took a model kit and follow the instructions, putting the things
Guido van Rossum (05:46.000)
together. I mean, I think the first few years that I built electronics kits, I really did
Guido van Rossum (05:51.680)
not have enough understanding of sort of electronics to really understand what I was doing. I mean,
Guido van Rossum (05:59.760)
I could debug it, and I could sort of follow the instructions very carefully, which has
Lex Fridman (06:06.480)
always stayed with me. But I had a very naive model of, like, how do I build a circuit?
Guido van Rossum (06:14.560)
Of, like, how a transistor works? And I don't think that in those days, I had any understanding
Guido van Rossum (06:22.800)
of coils and capacitors, which actually sort of was a major problem when I started to build
Guido van Rossum (06:32.560)
more complex digital circuits, because I was unaware of the sort of the analog part of
Guido van Rossum (06:39.840)
the – how they actually work. And I would have things that – the schematic looked
Guido van Rossum (06:50.080)
– everything looked fine, and it didn't work. And what I didn't realize was that
Guido van Rossum (06:57.440)
there was some megahertz level oscillation that was throwing the circuit off, because
Guido van Rossum (07:02.720)
I had a sort of – two wires were too close, or the switches were kind of poorly built.
Lex Fridman (07:13.360)
But through that time, I think it's really interesting and instructive to think about,
Guido van Rossum (07:19.280)
because echoes of it are in this time now. So in the 1970s, the personal computer was
Guido van Rossum (07:24.600)
being born. So did you sense, in tinkering with these circuits, did you sense the encroaching
Guido van Rossum (07:33.200)
revolution in personal computing? So if at that point, we would sit you down and ask
Guido van Rossum (07:39.320)
you to predict the 80s and the 90s, do you think you would be able to do so successfully
Guido van Rossum (07:46.040)
to unroll the process that's happening? No, I had no clue. I remember, I think, in
Guido van Rossum (07:55.560)
the summer after my senior year – or maybe it was the summer after my junior year – well,
Guido van Rossum (08:03.060)
at some point, I think, when I was 18, I went on a trip to the Math Olympiad in Eastern
Guido van Rossum (08:11.600)
Europe, and there was like – I was part of the Dutch team, and there were other nerdy
Guido van Rossum (08:16.920)
kids that sort of had different experiences, and one of them told me about this amazing
Guido van Rossum (08:23.040)
thing called a computer. And I had never heard that word. My own explorations in electronics
Guido van Rossum (08:31.840)
were sort of about very simple digital circuits, and I had sort of – I had the idea that
Guido van Rossum (08:40.420)
I somewhat understood how a digital calculator worked. And so there is maybe some echoes
Guido van Rossum (08:49.760)
of computers there, but I never made that connection. I didn't know that when my parents
Guido van Rossum (08:56.440)
were paying for magazine subscriptions using punched cards, that there was something called
Guido van Rossum (09:03.520)
a computer that was involved that read those cards and transferred the money between accounts.
Guido van Rossum (09:08.260)
I was also not really interested in those things. It was only when I went to university
Guido van Rossum (09:15.880)
to study math that I found out that they had a computer, and students were allowed to use
Lex Fridman (09:23.120)
it.
Lex Fridman (09:24.120)
And there were some – you're supposed to talk to that computer by programming it.
Lex Fridman (09:27.800)
What did that feel like, finding –
Guido van Rossum (09:29.920)
Yeah, that was the only thing you could do with it. The computer wasn't really connected
Guido van Rossum (09:35.440)
to the real world. The only thing you could do was sort of – you typed your program
Guido van Rossum (09:41.400)
on a bunch of punched cards. You gave the punched cards to the operator, and an hour
Guido van Rossum (09:47.840)
later the operator gave you back your printout. And so all you could do was write a program
Guido van Rossum (09:55.520)
that did something very abstract. And I don't even remember what my first forays into programming
Guido van Rossum (10:04.080)
were, but they were sort of doing simple math exercises and just to learn how a programming
Guido van Rossum (10:13.440)
language worked.
Guido van Rossum (10:15.560)
Did you sense, okay, first year of college, you see this computer, you're able to have
Guido van Rossum (10:21.680)
a program and it generates some output. Did you start seeing the possibility of this,
Guido van Rossum (10:29.420)
or was it a continuation of the tinkering with circuits? Did you start to imagine that
Guido van Rossum (10:34.920)
one, the personal computer, but did you see it as something that is a tool, like a word
Guido van Rossum (10:42.460)
processing tool, maybe for gaming or something? Or did you start to imagine that it could
Guido van Rossum (10:47.160)
be going to the world of robotics, like the Frankenstein picture that you could create
Guido van Rossum (10:53.860)
an artificial being? There's like another entity in front of you. You did not see the
Guido van Rossum (10:59.640)
computer.
Guido van Rossum (11:00.640)
I don't think I really saw it that way. I was really more interested in the tinkering.
Guido van Rossum (11:05.840)
It's maybe not a sort of a complete coincidence that I ended up sort of creating a programming
Guido van Rossum (11:14.920)
language which is a tool for other programmers. I've always been very focused on the sort
Guido van Rossum (11:20.360)
of activity of programming itself and not so much what happens with the program you
Lex Fridman (11:28.920)
write.
Guido van Rossum (11:29.920)
Right.
Guido van Rossum (11:30.920)
I do remember, and I don't remember, maybe in my second or third year, probably my second
Guido van Rossum (11:37.800)
actually, someone pointed out to me that there was this thing called Conway's Game of Life.
Lex Fridman (11:46.680)
You're probably familiar with it. I think –
Guido van Rossum (11:50.480)
In the 70s, I think is when they came up with it.
Lex Fridman (11:53.200)
So there was a Scientific American column by someone who did a monthly column about
Guido van Rossum (12:00.840)
mathematical diversions. I'm also blanking out on the guy's name. It was very famous
Guido van Rossum (12:06.580)
at the time and I think up to the 90s or so. And one of his columns was about Conway's
Guido van Rossum (12:12.440)
Game of Life and he had some illustrations and he wrote down all the rules and sort of
Guido van Rossum (12:18.160)
there was the suggestion that this was philosophically interesting, that that was why Conway had
Guido van Rossum (12:23.720)
called it that. And all I had was like the two pages photocopy of that article. I don't
Guido van Rossum (12:31.480)
even remember where I got it. But it spoke to me and I remember implementing a version
Guido van Rossum (12:40.200)
of that game for the batch computer we were using where I had a whole Pascal program that
Guido van Rossum (12:49.000)
sort of read an initial situation from input and read some numbers that said do so many
Guido van Rossum (12:56.480)
generations and print every so many generations and then out would come pages and pages of
Lex Fridman (13:05.960)
sort of things.
Guido van Rossum (13:08.480)
I remember much later I've done a similar thing using Python but that original version
Guido van Rossum (13:18.360)
I wrote at the time I found interesting because I combined it with some trick I had learned
Guido van Rossum (13:27.700)
during my electronics hobbyist times. I essentially first on paper I designed a simple circuit
Guido van Rossum (13:36.000)
built out of logic gates that took nine bits of input which is sort of the cell and its
Guido van Rossum (13:45.780)
neighbors and produced a new value for that cell and it's like a combination of a half
Guido van Rossum (13:54.040)
adder and some other clipping. It's actually a full adder. And so I had worked that out
Lex Fridman (14:01.040)
and then I translated that into a series of Boolean operations on Pascal integers where
Guido van Rossum (14:10.520)
you could use the integers as bitwise values. And so I could basically generate 60 bits
Guido van Rossum (14:21.740)
of a generation in like eight instructions or so.
Lex Fridman (14:28.800)
Nice.
Lex Fridman (14:29.800)
So I was proud of that.
Guido van Rossum (14:32.560)
It's funny that you mentioned, so for people who don't know Conway's Game of Life, it's
Guido van Rossum (14:38.120)
a cellular automata where there's single compute units that kind of look at their neighbors
Lex Fridman (14:44.840)
and figure out what they look like in the next generation based on the state of their
Guido van Rossum (14:50.080)
neighbors and this is deeply distributed system in concept at least. And then there's simple
Guido van Rossum (14:57.840)
rules that all of them follow and somehow out of this simple rule when you step back
Lex Fridman (15:04.400)
and look at what occurs, it's beautiful. There's an emergent complexity. Even though the underlying
Guido van Rossum (15:13.160)
rules are simple, there's an emergent complexity. Now the funny thing is you've implemented
Guido van Rossum (15:17.440)
this and the thing you're commenting on is you're proud of a hack you did to make it
Guido van Rossum (15:23.660)
run efficiently. When you're not commenting on, it's a beautiful implementation, you're
Guido van Rossum (15:30.800)
not commenting on the fact that there's an emergent complexity that you've coded a simple
Guido van Rossum (15:36.780)
program and when you step back and you print out the following generation after generation,
Guido van Rossum (15:42.960)
that's stuff that you may have not predicted would happen is happening.
Lex Fridman (15:48.400)
And is that magic? I mean, that's the magic that all of us feel when we program. When
Guido van Rossum (15:53.600)
you create a program and then you run it and whether it's Hello World or it shows something
Guido van Rossum (15:59.240)
on screen, if there's a graphical component, are you seeing the magic in the mechanism
Lex Fridman (16:03.840)
of creating that?
Guido van Rossum (16:05.200)
I think I went back and forth. As a student, we had an incredibly small budget of computer
Guido van Rossum (16:14.440)
time that we could use. It was actually measured. I once got in trouble with one of my professors
Lex Fridman (16:20.280)
because I had overspent the department's budget. It's a different story.
Guido van Rossum (16:29.640)
I actually wanted the efficient implementation because I also wanted to explore what would
Guido van Rossum (16:36.900)
happen with a larger number of generations and a larger size of the board. Once the implementation
Guido van Rossum (16:48.560)
was flawless, I would feed it different patterns and then I think maybe there was a follow
Guido van Rossum (16:57.000)
up article where there were patterns that were like gliders, patterns that repeated
Guido van Rossum (17:03.620)
themselves after a number of generations but translated one or two positions to the right
Guido van Rossum (17:13.200)
or up or something like that. I remember things like glider guns. Well, you can Google Conway's
Guido van Rossum (17:21.720)
Game of Life. People still go aww and ooh over it.
Guido van Rossum (17:27.560)
For a reason because it's not really well understood why. I mean, this is what Stephen
Guido van Rossum (17:32.680)
Wolfram is obsessed about. We don't have the mathematical tools to describe the kind of
Guido van Rossum (17:40.240)
complexity that emerges in these kinds of systems. The only way you can do is to run
Guido van Rossum (17:45.120)
it.
Guido van Rossum (17:47.120)
I'm not convinced that it's sort of a problem that lends itself to classic mathematical
Guido van Rossum (17:55.720)
analysis.
Guido van Rossum (17:59.920)
One theory of how you create an artificial intelligence or artificial being is you kind
Guido van Rossum (18:05.120)
of have to, same with the Game of Life, you kind of have to create a universe and let
Guido van Rossum (18:10.120)
it run. That creating it from scratch in a design way, coding up a Python program that
Guido van Rossum (18:17.520)
creates a fully intelligent system may be quite challenging. You might need to create
Lex Fridman (18:22.760)
a universe just like the Game of Life.
Guido van Rossum (18:27.120)
You might have to experiment with a lot of different universes before there is a set
Guido van Rossum (18:33.200)
of rules that doesn't essentially always just end up repeating itself in a trivial
Guido van Rossum (18:41.480)
way.
Guido van Rossum (18:42.480)
Yeah, and Stephen Wolfram works with these simple rules, says that it's kind of surprising
Lex Fridman (18:49.840)
how quickly you find rules that create interesting things. You shouldn't be able to, but somehow
Guido van Rossum (18:55.280)
you do. And so maybe our universe is laden with rules that will create interesting things
Guido van Rossum (19:02.120)
that might not look like humans, but emergent phenomena that's interesting may not be as
Lex Fridman (19:07.440)
difficult to create as we think.
Guido van Rossum (19:09.440)
Sure.
Lex Fridman (19:10.440)
But let me sort of ask, at that time, some of the world, at least in popular press, was
Guido van Rossum (19:17.440)
kind of captivated, perhaps at least in America, by the idea of artificial intelligence, that
Guido van Rossum (19:25.680)
these computers would be able to think pretty soon. And did that touch you at all? In science
Lex Fridman (19:33.240)
fiction or in reality in any way?
Guido van Rossum (19:37.800)
I didn't really start reading science fiction until much, much later. I think as a teenager
Guido van Rossum (19:49.000)
I read maybe one bundle of science fiction stories.
Lex Fridman (19:54.520)
Was it in the background somewhere, like in your thoughts?
Guido van Rossum (19:57.960)
That sort of the using computers to build something intelligent always felt to me, because
Lex Fridman (1:00:00.120)
And how did you try to design it into the language?
Guido van Rossum (1:00:03.560)
There are different tasks and as a programmer, it's useful to have different tools available
Lex Fridman (1:00:10.400)
that sort of are suitable for different tasks.
Lex Fridman (1:00:13.940)
So I still write C code, I still write shell code, but I write most of my things in Python.
Lex Fridman (1:00:25.600)
Why do I still use those other languages, because sometimes the task just demands it.
Lex Fridman (1:00:33.000)
And well, I would say most of the time the task actually demands a certain language because
Guido van Rossum (1:00:39.000)
the task is not write a program that solves problem X from scratch, but it's more like
Guido van Rossum (1:00:45.600)
fix a bug in existing program X or add a small feature to an existing large program.
Lex Fridman (1:00:56.680)
But even if you're not constrained in your choice of language by context like that, there
Guido van Rossum (1:01:10.160)
is still the fact that if you write it in a certain language, then you have this balance
Lex Fridman (1:01:21.360)
between how long does it take you to write the code and how long does the code run?
Lex Fridman (1:01:31.840)
And when you're in the phase of exploring solutions, you often spend much more time
Guido van Rossum (1:01:42.760)
writing the code than running it because every time you've run it, you see that the output
Guido van Rossum (1:01:50.720)
is not quite what you wanted and you spend some more time coding.
Lex Fridman (1:01:58.480)
And a language like Python just makes that iteration much faster because there are fewer
Guido van Rossum (1:02:06.760)
details that you have to get right before your program compiles and runs.
Guido van Rossum (1:02:19.480)
There are libraries that do all sorts of stuff for you, so you can sort of very quickly take
Guido van Rossum (1:02:26.400)
a bunch of existing components, put them together, and get your prototype application running.
Guido van Rossum (1:02:36.320)
Just like when I was building electronics, I was using a breadboard most of the time,
Lex Fridman (1:02:42.860)
so I had this sprawl out circuit that if you shook it, it would stop working because it
Guido van Rossum (1:02:51.320)
was not put together very well, but it functioned and all I wanted was to see that it worked
Lex Fridman (1:02:58.800)
and then move on to the next schematic or design or add something to it.
Guido van Rossum (1:03:05.000)
Once you've sort of figured out, oh, this is the perfect design for my radio or light
Lex Fridman (1:03:10.500)
sensor or whatever, then you can say, okay, how do we design a PCB for this?
Lex Fridman (1:03:15.800)
How do we solder the components in a small space?
Lex Fridman (1:03:19.920)
How do we make it so that it is robust against, say, voltage fluctuations or mechanical disruption?
Guido van Rossum (1:03:32.840)
I know nothing about that when it comes to designing electronics, but I know a lot about
Guido van Rossum (1:03:37.320)
that when it comes to writing code.
Lex Fridman (1:03:40.400)
So the initial steps are efficient, fast, and there's not much stuff that gets in the
Lex Fridman (1:03:46.080)
way, but you're kind of describing, like Darwin described the evolution of species, right?
Lex Fridman (1:03:56.680)
You're observing of what is true about Python.
Guido van Rossum (1:04:00.520)
Now if you take a step back, if the act of creating languages is art and you had three
Guido van Rossum (1:04:07.800)
months to do it, initial steps, so you just specified a bunch of goals, sort of things
Guido van Rossum (1:04:15.640)
that you observe about Python, perhaps you had those goals, but how do you create the
Lex Fridman (1:04:19.400)
rules, the syntactic structure, the features that result in those?
Lex Fridman (1:04:25.600)
So I have in the beginning and I have follow up questions about through the evolution of
Guido van Rossum (1:04:29.880)
Python too, but in the very beginning when you were sitting there creating the lexical
Guido van Rossum (1:04:35.440)
analyzer or whatever.
Guido van Rossum (1:04:37.440)
Python was still a big part of it because I sort of, I said to myself, I don't want
Guido van Rossum (1:04:47.240)
to have to design everything from scratch, I'm going to borrow features from other languages
Lex Fridman (1:04:53.640)
that I like.
Guido van Rossum (1:04:54.640)
Oh, interesting.
Lex Fridman (1:04:55.640)
So you basically, exactly, you first observe what you like.
Guido van Rossum (1:04:58.360)
Yeah, and so that's why if you're 17 years old and you want to sort of create a programming
Guido van Rossum (1:05:05.240)
language, you're not going to be very successful at it because you have no experience with
Guido van Rossum (1:05:11.600)
other languages, whereas I was in my, let's say mid 30s, I had written parsers before,
Lex Fridman (1:05:24.300)
so I had worked on the implementation of ABC, I had spent years debating the design of ABC
Guido van Rossum (1:05:30.880)
with its authors, with its designers, I had nothing to do with the design, it was designed
Lex Fridman (1:05:37.520)
fully as it ended up being implemented when I joined the team.
Lex Fridman (1:05:42.080)
But so you borrow ideas and concepts and very concrete sort of local rules from different
Guido van Rossum (1:05:51.440)
languages like the indentation and certain other syntactic features from ABC, but I chose
Guido van Rossum (1:05:58.920)
to borrow string literals and how numbers work from C and various other things.
Lex Fridman (1:06:07.960)
So in then, if you take that further, so yet you've had this funny sounding, but I think
Guido van Rossum (1:06:13.800)
surprisingly accurate and at least practical title of benevolent dictator for life for
Guido van Rossum (1:06:21.000)
quite, you know, for the last three decades or whatever, or no, not the actual title,
Lex Fridman (1:06:25.240)
but functionally speaking.
Lex Fridman (1:06:27.940)
So you had to make decisions, design decisions.
Lex Fridman (1:06:34.280)
Can you maybe, let's take Python 2, so releasing Python 3 as an example.
Lex Fridman (1:06:41.960)
It's not backward compatible to Python 2 in ways that a lot of people know.
Lex Fridman (1:06:47.240)
So what was that deliberation, discussion, decision like?
Lex Fridman (1:06:50.640)
Yeah.
Lex Fridman (1:06:51.640)
What was the psychology of that experience?
Lex Fridman (1:06:54.520)
Do you regret any aspects of how that experience undergone that?
Guido van Rossum (1:06:58.520)
Well, yeah, so it was a group process really.
Guido van Rossum (1:07:03.040)
At that point, even though I was BDFL in name and certainly everybody sort of respected
Guido van Rossum (1:07:11.880)
my position as the creator and the current sort of owner of the language design, I was
Lex Fridman (1:07:22.160)
looking at everyone else for feedback.
Guido van Rossum (1:07:26.560)
Sort of Python 3.0 in some sense was sparked by other people in the community pointing
Lex Fridman (1:07:35.280)
out, oh, well, there are a few issues that sort of bite users over and over.
Lex Fridman (1:07:46.360)
Can we do something about that?
Lex Fridman (1:07:48.920)
And for Python 3, we took a number of those Python words as they were called at the time
Lex Fridman (1:07:56.360)
and we said, can we try to sort of make small changes to the language that address those
Lex Fridman (1:08:04.800)
words?
Lex Fridman (1:08:06.560)
And we had sort of in the past, we had always taken backwards compatibility very seriously.
Lex Fridman (1:08:15.360)
And so many Python words in earlier versions had already been resolved because they could
Guido van Rossum (1:08:20.420)
be resolved while maintaining backwards compatibility or sort of using a very gradual path of evolution
Lex Fridman (1:08:29.740)
of the language in a certain area.
Lex Fridman (1:08:31.960)
And so we were stuck with a number of words that were widely recognized as problems, not
Guido van Rossum (1:08:39.760)
like roadblocks, but nevertheless sort of things that some people trip over and you know that
Guido van Rossum (1:08:47.680)
that's always the same thing that people trip over when they trip.
Lex Fridman (1:08:52.080)
And we could not think of a backwards compatible way of resolving those issues.
Lex Fridman (1:08:58.480)
But it's still an option to not resolve the issues, right?
Lex Fridman (1:09:01.920)
And so yes, for a long time, we had sort of resigned ourselves to, well, okay, the language
Guido van Rossum (1:09:07.920)
is not going to be perfect in this way and that way and that way.
Lex Fridman (1:09:13.400)
And we sort of, certain of these, I mean, there are still plenty of things where you
Guido van Rossum (1:09:19.440)
can say, well, that particular detail is better in Java or in R or in Visual Basic or whatever.
Lex Fridman (1:09:32.680)
And we're okay with that because, well, we can't easily change it.
Guido van Rossum (1:09:37.960)
It's not too bad.
Guido van Rossum (1:09:38.960)
We can do a little bit with user education or we can have a static analyzer or warnings
Guido van Rossum (1:09:47.180)
in the parse or something.
Lex Fridman (1:09:49.440)
But there were things where we thought, well, these are really problems that are not going
Guido van Rossum (1:09:54.880)
away.
Lex Fridman (1:09:55.880)
They are getting worse in the future.
Guido van Rossum (1:10:00.840)
We should do something about that.
Lex Fridman (1:10:03.040)
But ultimately there is a decision to be made, right?
Lex Fridman (1:10:05.640)
So was that the toughest decision in the history of Python you had to make as the benevolent
Lex Fridman (1:10:13.320)
dictator for life?
Guido van Rossum (1:10:15.180)
Or if not, what are there, maybe even on the smaller scale, what was the decision where
Lex Fridman (1:10:20.160)
you were really torn up about?
Guido van Rossum (1:10:22.040)
Well, the toughest decision was probably to resign.
Lex Fridman (1:10:25.800)
All right, let's go there.
Guido van Rossum (1:10:28.120)
Hold on a second then.
Guido van Rossum (1:10:29.360)
Let me just, because in the interest of time too, because I have a few cool questions for
Guido van Rossum (1:10:33.200)
you and let's touch a really important one because it was quite dramatic and beautiful
Lex Fridman (1:10:38.160)
in certain kinds of ways.
Guido van Rossum (1:10:40.400)
In July this year, three months ago, you wrote, now that PEP 572 is done, I don't ever want
Guido van Rossum (1:10:47.320)
to have to fight so hard for a PEP and find that so many people despise my decisions.
Guido van Rossum (1:10:52.680)
I would like to remove myself entirely from the decision process.
Guido van Rossum (1:10:56.240)
I'll still be there for a while as an ordinary core developer and I'll still be available
Guido van Rossum (1:11:01.520)
to mentor people, possibly more available.
Lex Fridman (1:11:05.440)
But I'm basically giving myself a permanent vacation from being BDFL, benevolent dictator
Guido van Rossum (1:11:11.000)
for life.
Lex Fridman (1:11:12.000)
And you all will be on your own.
Guido van Rossum (1:11:14.240)
First of all, it's almost Shakespearean.
Lex Fridman (1:11:19.720)
I'm not going to appoint a successor.
Lex Fridman (1:11:22.300)
So what are you all going to do?
Lex Fridman (1:11:24.640)
Create a democracy, anarchy, a dictatorship, a federation?
Lex Fridman (1:11:29.240)
So that was a very dramatic and beautiful set of statements.
Guido van Rossum (1:11:34.560)
It's almost, it's open ended nature called the community to create a future for Python.
Guido van Rossum (1:11:40.080)
It's just kind of a beautiful aspect to it.
Lex Fridman (1:11:43.280)
So what, and dramatic, you know, what was making that decision like?
Lex Fridman (1:11:48.320)
What was on your heart, on your mind, stepping back now a few months later?
Lex Fridman (1:11:54.560)
I'm glad you liked the writing because it was actually written pretty quickly.
Guido van Rossum (1:12:02.940)
It was literally something like after months and months of going around in circles, I had
Guido van Rossum (1:12:14.240)
finally approved PEP572, which I had a big hand in its design, although I didn't initiate
Guido van Rossum (1:12:26.240)
it originally.
Guido van Rossum (1:12:27.760)
I sort of gave it a bunch of nudges in a direction that would be better for the language.
Lex Fridman (1:12:36.320)
So sorry, just to ask, is async IO, that's the one or no?
Lex Fridman (1:12:40.320)
PEP572 was actually a small feature, which is assignment expressions.
Guido van Rossum (1:12:49.320)
That had been, there was just a lot of debate where a lot of people claimed that they knew
Lex Fridman (1:12:58.200)
what was Pythonic and what was not Pythonic, and they knew that this was going to destroy
Guido van Rossum (1:13:04.800)
the language.
Guido van Rossum (1:13:06.080)
This was like a violation of Python's most fundamental design philosophy, and I thought
Guido van Rossum (1:13:11.800)
that was all bullshit because I was in favor of it, and I would think I know something
Lex Fridman (1:13:17.200)
about Python's design philosophy.
Lex Fridman (1:13:19.120)
So I was really tired and also stressed of that thing, and literally after sort of announcing
Guido van Rossum (1:13:26.340)
I was going to accept it, a certain Wednesday evening I had finally sent the email, it's
Guido van Rossum (1:13:34.560)
accepted.
Lex Fridman (1:13:35.560)
I can just go implement it.
Lex Fridman (1:13:38.920)
So I went to bed feeling really relieved, that's behind me.
Lex Fridman (1:13:44.120)
And I wake up Thursday morning, 7 a.m., and I think, well, that was the last one that's
Guido van Rossum (1:13:54.320)
going to be such a terrible debate, and that's the last time that I let myself be so stressed
Lex Fridman (1:14:03.880)
out about a pep decision.
Guido van Rossum (1:14:06.520)
I should just resign.
Guido van Rossum (1:14:07.920)
I've been sort of thinking about retirement for half a decade, I've been joking and sort
Guido van Rossum (1:14:15.520)
of mentioning retirement, sort of telling the community at some point in the future
Lex Fridman (1:14:22.460)
I'm going to retire, don't take that FL part of my title too literally.
Lex Fridman (1:14:29.400)
And I thought, okay, this is it.
Guido van Rossum (1:14:32.080)
I'm done, I had the day off, I wanted to have a good time with my wife, we were going to
Guido van Rossum (1:14:39.200)
a little beach town nearby, and in I think maybe 15, 20 minutes I wrote that thing that
Lex Fridman (1:14:48.480)
you just called Shakespearean.
Guido van Rossum (1:14:51.320)
The funny thing is I didn't even realize what a monumental decision it was, because
Guido van Rossum (1:15:01.560)
five minutes later I read that link to my message back on Twitter, where people were
Guido van Rossum (1:15:09.200)
already discussing on Twitter, Guido resigned as the BDFL.
Lex Fridman (1:15:15.280)
And I had posted it on an internal forum that I thought was only read by core developers,
Lex Fridman (1:15:22.440)
so I thought I would at least have one day before the news would sort of get out.
Guido van Rossum (1:15:28.520)
The on your own aspects had also an element of quite, it was quite a powerful element
Guido van Rossum (1:15:36.200)
of the uncertainty that lies ahead, but can you also just briefly talk about, for example
Guido van Rossum (1:15:43.080)
I play guitar as a hobby for fun, and whenever I play people are super positive, super friendly,
Guido van Rossum (1:15:49.920)
they're like, this is awesome, this is great.
Lex Fridman (1:15:52.680)
But sometimes I enter as an outside observer, I enter the programming community and there
Guido van Rossum (1:15:57.520)
seems to sometimes be camps on whatever the topic, and the two camps, the two or plus
Lex Fridman (1:16:05.560)
camps, are often pretty harsh at criticizing the opposing camps.
Lex Fridman (1:16:11.700)
As an onlooker, I may be totally wrong on this, but what do you think of this?
Lex Fridman (1:16:14.880)
Yeah, holy wars are sort of a favorite activity in the programming community.
Lex Fridman (1:16:19.760)
And what is the psychology behind that?
Lex Fridman (1:16:22.120)
Is that okay for a healthy community to have?
Lex Fridman (1:16:25.120)
Is that a productive force ultimately for the evolution of a language?
Guido van Rossum (1:16:29.760)
Well, if everybody is patting each other on the back and never telling the truth, it would
Guido van Rossum (1:16:39.080)
not be a good thing.
Guido van Rossum (1:16:40.840)
I think there is a middle ground where sort of being nasty to each other is not okay,
Lex Fridman (1:16:52.760)
but there is a middle ground where there is healthy ongoing criticism and feedback that
Lex Fridman (1:17:01.760)
is very productive.
Lex Fridman (1:17:04.780)
And you mean at every level you see that.
Guido van Rossum (1:17:07.760)
I mean, someone proposes to fix a very small issue in a code base, chances are that some
Guido van Rossum (1:17:17.760)
reviewer will sort of respond by saying, well, actually, you can do it better the other way.
Guido van Rossum (1:17:27.080)
When it comes to deciding on the future of the Python core developer community, we now
Guido van Rossum (1:17:34.360)
have, I think, five or six competing proposals for a constitution.
Lex Fridman (1:17:41.160)
So that future, do you have a fear of that future, do you have a hope for that future?
Guido van Rossum (1:17:48.040)
I'm very confident about that future.
Lex Fridman (1:17:51.280)
By and large, I think that the debate has been very healthy and productive.
Lex Fridman (1:17:58.920)
And I actually, when I wrote that resignation email, I knew that Python was in a very good
Guido van Rossum (1:18:07.680)
spot and that the Python core developer community, the group of 50 or 100 people who sort of
Guido van Rossum (1:18:16.840)
write or review most of the code that goes into Python, those people get along very well
Lex Fridman (1:18:24.720)
most of the time.
Guido van Rossum (1:18:27.680)
A large number of different areas of expertise are represented, different levels of experience
Guido van Rossum (1:18:40.120)
in the Python core dev community, different levels of experience completely outside it
Guido van Rossum (1:18:45.440)
in software development in general, large systems, small systems, embedded systems.
Lex Fridman (1:18:53.040)
So I felt okay resigning because I knew that the community can really take care of itself.
Lex Fridman (1:19:03.880)
And out of a grab bag of future feature developments, let me ask if you can comment, maybe on all
Lex Fridman (1:19:12.360)
very quickly, concurrent programming, parallel computing, async IO.
Guido van Rossum (1:19:19.120)
These are things that people have expressed hope, complained about, whatever, have discussed
Lex Fridman (1:19:24.880)
on Reddit.
Guido van Rossum (1:19:25.880)
Async IO, so the parallelization in general, packaging, I was totally clueless on this.
Guido van Rossum (1:19:32.200)
I just used pip to install stuff, but apparently there's pipenv, poetry, there's these dependency
Guido van Rossum (1:19:38.600)
packaging systems that manage dependencies and so on.
Guido van Rossum (1:19:41.300)
They're emerging and there's a lot of confusion about what's the right thing to use.
Guido van Rossum (1:19:45.520)
Then also functional programming, are we going to get more functional programming or not,
Lex Fridman (1:19:56.360)
this kind of idea.
Lex Fridman (1:19:59.040)
And of course the GIL connected to the parallelization, I suppose, the global interpreter lock problem.
Lex Fridman (1:20:08.280)
Can you just comment on whichever you want to comment on?
Guido van Rossum (1:20:12.800)
Well, let's take the GIL and parallelization and async IO as one topic.
Guido van Rossum (1:20:25.440)
I'm not that hopeful that Python will develop into a sort of high concurrency, high parallelism
Guido van Rossum (1:20:35.820)
language.
Guido van Rossum (1:20:37.960)
That's sort of the way the language is designed, the way most users use the language, the way
Guido van Rossum (1:20:44.800)
the language is implemented, all make that a pretty unlikely future.
Lex Fridman (1:20:50.280)
So you think it might not even need to, really the way people use it, it might not be something
Guido van Rossum (1:20:56.040)
that should be of great concern.
Guido van Rossum (1:20:58.160)
I think async IO is a special case because it sort of allows overlapping IO and only
Guido van Rossum (1:21:05.620)
IO and that is a sort of best practice of supporting very high throughput IO, many connections
Lex Fridman (1:21:18.160)
per second.
Guido van Rossum (1:21:21.680)
I'm not worried about that.
Lex Fridman (1:21:22.780)
I think async IO will evolve.
Guido van Rossum (1:21:25.280)
There are a couple of competing packages.
Guido van Rossum (1:21:27.440)
We have some very smart people who are sort of pushing us to make async IO better.
Guido van Rossum (1:21:36.800)
Parallel computing, I think that Python is not the language for that.
Guido van Rossum (1:21:43.800)
There are ways to work around it, but you can't expect to write an algorithm in Python
Lex Fridman (1:21:53.560)
and have a compiler automatically parallelize that.
Lex Fridman (1:21:57.440)
What you can do is use a package like NumPy and there are a bunch of other very powerful
Guido van Rossum (1:22:03.520)
packages that sort of use all the CPUs available because you tell the package, here's the data,
Guido van Rossum (1:22:12.480)
here's the abstract operation to apply over it, go at it, and then we're back in the C++
Guido van Rossum (1:22:19.040)
world.
Lex Fridman (1:22:20.040)
Those packages are themselves implemented usually in C++.
Guido van Rossum (1:22:24.600)
That's where TensorFlow and all these packages come in, where they parallelize across GPUs,
Lex Fridman (1:22:28.000)
for example, they take care of that for you.
Lex Fridman (1:22:30.480)
In terms of packaging, can you comment on the future of packaging in Python?
Lex Fridman (1:22:36.600)
Packaging has always been my least favorite topic.
Guido van Rossum (1:22:42.640)
It's a really tough problem because the OS and the platform want to own packaging, but
Lex Fridman (1:22:55.600)
their packaging solution is not specific to a language.
Guido van Rossum (1:23:01.000)
If you take Linux, there are two competing packaging solutions for Linux or for Unix
Lex Fridman (1:23:07.480)
in general, but they all work across all languages.
Guido van Rossum (1:23:15.000)
Several languages like Node, JavaScript, Ruby, and Python all have their own packaging solutions
Lex Fridman (1:23:24.760)
that only work within the ecosystem of that language.
Lex Fridman (1:23:29.480)
What should you use?
Lex Fridman (1:23:31.920)
That is a tough problem.
Guido van Rossum (1:23:34.560)
My own approach is I use the system packaging system to install Python, and I use the Python
Lex Fridman (1:23:43.520)
packaging system then to install third party Python packages.
Guido van Rossum (1:23:49.280)
That's what most people do.
Lex Fridman (1:23:51.480)
Ten years ago, Python packaging was really a terrible situation.
Guido van Rossum (1:23:56.400)
Nowadays, pip is the future, there is a separate ecosystem for numerical and scientific Python
Lex Fridman (1:24:05.360)
based on Anaconda.
Guido van Rossum (1:24:08.200)
Those two can live together.
Lex Fridman (1:24:09.760)
I don't think there is a need for more than that.
Guido van Rossum (1:24:13.600)
That's packaging.
Lex Fridman (1:24:14.600)
Well, at least for me, that's where I've been extremely happy.
Guido van Rossum (1:24:18.720)
I didn't even know this was an issue until it was brought up.
Guido van Rossum (1:24:22.320)
In the interest of time, let me sort of skip through a million other questions I have.
Lex Fridman (1:24:27.600)
So I watched the five and a half hour oral history that you've done with the Computer
Guido van Rossum (1:24:32.880)
History Museum, and the nice thing about it, it gave this, because of the linear progression
Guido van Rossum (1:24:37.600)
of the interview, it gave this feeling of a life, you know, a life well lived with interesting
Guido van Rossum (1:24:44.480)
things in it, sort of a pretty, I would say a good spend of this little existence we have
Guido van Rossum (1:24:52.160)
on Earth.
Guido van Rossum (1:24:53.160)
So, outside of your family, looking back, what about this journey are you really proud
Lex Fridman (1:24:59.840)
of?
Lex Fridman (1:25:00.840)
Are there moments that stand out, accomplishments, ideas?
Guido van Rossum (1:25:07.040)
Is it the creation of Python itself that stands out as a thing that you look back and say,
Lex Fridman (1:25:14.040)
damn, I did pretty good there?
Guido van Rossum (1:25:16.480)
Well, I would say that Python is definitely the best thing I've ever done, and I wouldn't
Guido van Rossum (1:25:25.520)
sort of say just the creation of Python, but the way I sort of raised Python, like a baby.
Guido van Rossum (1:25:36.560)
I didn't just conceive a child, but I raised a child, and now I'm setting the child free
Guido van Rossum (1:25:42.480)
in the world, and I've set up the child to sort of be able to take care of himself, and
Guido van Rossum (1:25:50.200)
I'm very proud of that.
Lex Fridman (1:25:52.640)
And as the announcer of Monty Python's Flying Circus used to say, and now for something
Guido van Rossum (1:25:56.740)
completely different, do you have a favorite Monty Python moment, or a moment in Hitchhiker's
Lex Fridman (1:26:02.280)
Guide, or any other literature show or movie that cracks you up when you think about it?
Guido van Rossum (1:26:07.720)
You can always play me the dead parrot sketch.
Lex Fridman (1:26:11.320)
Oh, that's brilliant.
Guido van Rossum (1:26:13.680)
That's my favorite as well.
Lex Fridman (1:26:14.680)
It's pushing up the daisies.
Guido van Rossum (1:26:15.680)
Okay, Greta, thank you so much for talking with me today.
Lex Fridman (1:26:20.680)
Lex, this has been a great conversation.
Guido van Rossum (20:04.720)
I felt I had so much understanding of what actually goes on inside a computer. I knew
Lex Fridman (20:12.920)
how many bits of memory it had and how difficult it was to program. And sort of, I didn't believe
Guido van Rossum (20:22.880)
at all that you could just build something intelligent out of that, that would really
Guido van Rossum (20:30.560)
sort of satisfy my definition of intelligence. I think the most influential thing that I
Guido van Rossum (20:40.600)
read in my early twenties was Gödel Escherbach. That was about consciousness, and that was
Lex Fridman (20:48.680)
a big eye opener in some sense.
Guido van Rossum (20:54.040)
In what sense? So, on your own brain, did you at the time or do you now see your own
Guido van Rossum (21:00.760)
brain as a computer? Or is there a total separation of the way? So yeah, you're very pragmatically
Guido van Rossum (21:07.720)
practically know the limits of memory, the limits of this sequential computing or weakly
Guido van Rossum (21:14.600)
paralyzed computing, and you just know what we have now, and it's hard to see how it creates.
Lex Fridman (21:21.000)
But it's also easy to see, it was in the 40s, 50s, 60s, and now at least similarities between
Lex Fridman (21:29.920)
the brain and our computers.
Guido van Rossum (21:31.680)
Oh yeah, I mean, I totally believe that brains are computers in some sense. I mean, the rules
Guido van Rossum (21:43.200)
they use to play by are pretty different from the rules we can sort of implement in our
Guido van Rossum (21:51.200)
current hardware, but I don't believe in, like, a separate thing that infuses us with
Guido van Rossum (22:02.960)
intelligence or consciousness or any of that. There's no soul, I've been an atheist
Guido van Rossum (22:10.480)
probably from when I was 10 years old, just by thinking a bit about math and the universe,
Lex Fridman (22:18.800)
and well, my parents were atheists. Now, I know that you could be an atheist and still
Guido van Rossum (22:26.640)
believe that there is something sort of about intelligence or consciousness that cannot
Guido van Rossum (22:34.080)
possibly emerge from a fixed set of rules. I am not in that camp. I totally see that,
Guido van Rossum (22:44.560)
sort of, given how many millions of years evolution took its time, DNA is a particular
Guido van Rossum (22:53.680)
machine that sort of encodes information and an unlimited amount of information in chemical
Guido van Rossum (23:07.040)
form and has figured out a way to replicate itself.
Guido van Rossum (23:12.320)
I thought that that was, maybe it's 300 million years ago, but I thought it was closer
Guido van Rossum (23:16.880)
to half a billion years ago, that that's sort of originated and it hasn't really changed,
Guido van Rossum (23:25.120)
that the sort of the structure of DNA hasn't changed ever since. That is like our binary
Guido van Rossum (23:32.040)
code that we have in hardware. I mean...
Guido van Rossum (23:35.200)
The basic programming language hasn't changed, but maybe the programming itself...
Guido van Rossum (23:39.760)
Obviously, it did sort of, it happened to be a set of rules that was good enough to
Guido van Rossum (23:48.320)
sort of develop endless variability and sort of the idea of self replicating molecules
Guido van Rossum (23:59.520)
competing with each other for resources and one type eventually sort of always taking
Guido van Rossum (24:05.360)
over. That happened before there were any fossils, so we don't know how that exactly
Guido van Rossum (24:12.320)
happened, but I believe it's clear that that did happen.
Lex Fridman (24:17.920)
Can you comment on consciousness and how you see it? Because I think we'll talk about
Guido van Rossum (24:25.360)
programming quite a bit. We'll talk about, you know, intelligence connecting to programming
Guido van Rossum (24:30.080)
fundamentally, but consciousness is this whole other thing. Do you think about it often as
Lex Fridman (24:38.080)
a developer of a programming language and as a human?
Guido van Rossum (24:45.440)
Those are pretty sort of separate topics. Sort of my line of work working with programming
Guido van Rossum (24:55.000)
does not involve anything that goes in the direction of developing intelligence or consciousness,
Lex Fridman (25:02.720)
but sort of privately as an avid reader of popular science writing, I have some thoughts
Guido van Rossum (25:13.880)
which is mostly that I don't actually believe that consciousness is an all or nothing thing.
Guido van Rossum (25:25.680)
I have a feeling that, and I forget what I read that influenced this, but I feel that
Guido van Rossum (25:35.960)
if you look at a cat or a dog or a mouse, they have some form of intelligence. If you
Guido van Rossum (25:41.400)
look at a fish, it has some form of intelligence, and that evolution just took a long time,
Lex Fridman (25:54.040)
but I feel that the sort of evolution of more and more intelligence that led to sort of
Guido van Rossum (26:01.320)
the human form of intelligence followed the evolution of the senses, especially the visual
Guido van Rossum (26:12.920)
sense. I mean, there is an enormous amount of processing that's needed to interpret
Lex Fridman (26:20.480)
a scene, and humans are still better at that than computers are.
Lex Fridman (26:28.240)
And I have a feeling that there is a sort of, the reason that like mammals in particular
Guido van Rossum (26:39.660)
developed the levels of consciousness that they have and that eventually sort of going
Guido van Rossum (26:47.960)
from intelligence to self awareness and consciousness has to do with sort of being a robot that
Lex Fridman (26:55.360)
has very highly developed senses.
Guido van Rossum (26:58.920)
Has a lot of rich sensory information coming in, so that's a really interesting thought
Guido van Rossum (27:04.760)
that whatever that basic mechanism of DNA, whatever that basic building blocks of programming,
Guido van Rossum (27:14.200)
if you just add more abilities, more high resolution sensors, more sensors, you just
Guido van Rossum (27:21.080)
keep stacking those things on top that this basic programming in trying to survive develops
Guido van Rossum (27:26.760)
very interesting things that start to us humans to appear like intelligence and consciousness.
Guido van Rossum (27:35.000)
As far as robots go, I think that the self driving cars have that sort of the greatest
Guido van Rossum (27:42.280)
opportunity of developing something like that, because when I drive myself, I don't just
Lex Fridman (27:50.400)
pay attention to the rules of the road.
Guido van Rossum (27:53.800)
I also look around and I get clues from that, oh, this is a shopping district, oh, here's
Guido van Rossum (28:01.220)
an old lady crossing the street, oh, here is someone carrying a pile of mail, there's
Guido van Rossum (28:08.960)
a mailbox, I bet you they're going to cross the street to reach that mailbox.
Lex Fridman (28:14.040)
And I slow down, and I don't even think about that.
Lex Fridman (28:17.520)
And so, there is so much where you turn your observations into an understanding of what
Guido van Rossum (28:25.780)
other consciousnesses are going to do, or what other systems in the world are going
Guido van Rossum (28:32.680)
to be, oh, that tree is going to fall.
Guido van Rossum (28:37.400)
I see sort of, I see much more of, I expect somehow that if anything is going to become
Guido van Rossum (28:46.800)
unconscious, it's going to be the self driving car and not the network of a bazillion computers
Guido van Rossum (28:55.520)
in a Google or Amazon data center that are all networked together to do whatever they
Guido van Rossum (29:03.160)
do.
Guido van Rossum (29:04.160)
So, in that sense, so you actually highlight, because that's what I work in Thomas Vehicles,
Guido van Rossum (29:09.640)
you highlight the big gap between what we currently can't do and what we truly need
Lex Fridman (29:15.600)
to be able to do to solve the problem.
Guido van Rossum (29:18.500)
Under that formulation, then consciousness and intelligence is something that basically
Guido van Rossum (29:24.600)
a system should have in order to interact with us humans, as opposed to some kind of
Guido van Rossum (29:30.020)
abstract notion of a consciousness.
Guido van Rossum (29:35.280)
Consciousness is something that you need to have to be able to empathize, to be able to
Guido van Rossum (29:39.200)
fear, understand what the fear of death is, all these aspects that are important for interacting
Guido van Rossum (29:47.440)
with pedestrians, you need to be able to do basic computation based on our human desires
Lex Fridman (29:56.160)
and thoughts.
Lex Fridman (29:57.160)
And if you sort of, yeah, if you look at the dog, the dog clearly knows, I mean, I'm
Guido van Rossum (30:02.280)
not the dog owner, but I have friends who have dogs, the dogs clearly know what the
Guido van Rossum (30:07.340)
humans around them are going to do, or at least they have a model of what those humans
Guido van Rossum (30:11.400)
are going to do and they learn.
Guido van Rossum (30:14.160)
Some dogs know when you're going out and they want to go out with you, they're sad when
Guido van Rossum (30:19.060)
you leave them alone, they cry, they're afraid because they were mistreated when they were
Lex Fridman (30:26.080)
younger.
Guido van Rossum (30:31.040)
We don't assign sort of consciousness to dogs, or at least not all that much, but I also
Lex Fridman (30:39.280)
don't think they have none of that.
Lex Fridman (30:42.500)
So I think it's consciousness and intelligence are not all or nothing.
Lex Fridman (30:50.360)
The spectrum is really interesting.
Lex Fridman (30:52.780)
But in returning to programming languages and the way we think about building these
Guido van Rossum (30:58.760)
kinds of things, about building intelligence, building consciousness, building artificial
Guido van Rossum (31:03.260)
beings.
Lex Fridman (31:04.260)
So I think one of the exciting ideas came in the 17th century and with Leibniz, Hobbes,
Guido van Rossum (31:10.920)
Descartes, where there's this feeling that you can convert all thought, all reasoning,
Guido van Rossum (31:18.520)
all the thing that we find very special in our brains, you can convert all of that into
Guido van Rossum (31:24.480)
logic.
Lex Fridman (31:25.480)
So you can formalize it, formal reasoning, and then once you formalize everything, all
Guido van Rossum (31:30.400)
of knowledge, then you can just calculate and that's what we're doing with our brains
Lex Fridman (31:34.400)
is we're calculating.
Lex Fridman (31:35.400)
So there's this whole idea that this is possible, that this we can actually program.
Lex Fridman (31:40.240)
But they weren't aware of the concept of pattern matching in the sense that we are aware of
Guido van Rossum (31:46.520)
it now.
Guido van Rossum (31:47.640)
They sort of thought they had discovered incredible bits of mathematics like Newton's calculus
Lex Fridman (31:57.640)
and their sort of idealism, their sort of extension of what they could do with logic
Lex Fridman (32:06.840)
and math sort of went along those lines and they thought there's like, yeah, logic.
Guido van Rossum (32:18.000)
There's like a bunch of rules and a bunch of input.
Guido van Rossum (32:22.020)
They didn't realize that how you recognize a face is not just a bunch of rules but is
Guido van Rossum (32:28.600)
a shit ton of data plus a circuit that sort of interprets the visual clues and the context
Lex Fridman (32:39.160)
and everything else and somehow can massively parallel pattern match against stored rules.
Guido van Rossum (32:49.400)
I mean, if I see you tomorrow here in front of the Dropbox office, I might recognize you.
Guido van Rossum (32:56.320)
Even if I'm wearing a different shirt, yeah, but if I see you tomorrow in a coffee shop
Guido van Rossum (33:01.320)
in Belmont, I might have no idea that it was you or on the beach or whatever.
Lex Fridman (33:06.640)
I make those kind of mistakes myself all the time.
Guido van Rossum (33:10.160)
I see someone that I only know as like, oh, this person is a colleague of my wife's and
Lex Fridman (33:16.320)
then I see them at the movies and I didn't recognize them.
Lex Fridman (33:20.860)
But do you see those, you call it pattern matching, do you see that rules is unable
Lex Fridman (33:29.320)
to encode that?
Guido van Rossum (33:32.380)
Everything you see, all the pieces of information you look around this room, I'm wearing a black
Guido van Rossum (33:36.320)
shirt, I have a certain height, I'm a human, all these, there's probably tens of thousands
Guido van Rossum (33:41.720)
of facts you pick up moment by moment about this scene.
Guido van Rossum (33:45.680)
You take them for granted and you aggregate them together to understand the scene.
Guido van Rossum (33:50.000)
You don't think all of that could be encoded to where at the end of the day, you can just
Lex Fridman (33:53.800)
put it all on the table and calculate?
Guido van Rossum (33:57.440)
I don't know what that means.
Guido van Rossum (33:58.840)
I mean, yes, in the sense that there is no actual magic there, but there are enough layers
Guido van Rossum (34:08.680)
of abstraction from the facts as they enter my eyes and my ears to the understanding of
Guido van Rossum (34:17.640)
the scene that I don't think that AI has really covered enough of that distance.
Guido van Rossum (34:29.240)
It's like if you take a human body and you realize it's built out of atoms, well, that
Lex Fridman (34:37.800)
is a uselessly reductionist view, right?
Guido van Rossum (34:41.960)
The body is built out of organs, the organs are built out of cells, the cells are built
Guido van Rossum (34:46.380)
out of proteins, the proteins are built out of amino acids, the amino acids are built
Guido van Rossum (34:53.240)
out of atoms and then you get to quantum mechanics.
Lex Fridman (34:58.040)
So that's a very pragmatic view.
Guido van Rossum (34:59.920)
I mean, obviously as an engineer, I agree with that kind of view, but you also have
Guido van Rossum (35:03.720)
to consider the Sam Harris view of, well, intelligence is just information processing.
Guido van Rossum (35:13.160)
Like you said, you take in sensory information, you do some stuff with it and you come up
Lex Fridman (35:17.320)
with actions that are intelligent.
Guido van Rossum (35:20.760)
That makes it sound so easy.
Lex Fridman (35:22.480)
I don't know who Sam Harris is.
Guido van Rossum (35:24.280)
Oh, well, it's a philosopher.
Lex Fridman (35:26.400)
So like this is how philosophers often think, right?
Lex Fridman (35:29.680)
And essentially that's what Descartes was, is wait a minute, if there is, like you said,
Guido van Rossum (35:33.760)
no magic, so he basically says it doesn't appear like there's any magic, but we know
Lex Fridman (35:39.320)
so little about it that it might as well be magic.
Lex Fridman (35:44.280)
So just because we know that we're made of atoms, just because we know we're made
Guido van Rossum (35:47.800)
of organs, the fact that we know very little how to get from the atoms to organs in a way
Guido van Rossum (35:53.280)
that's recreatable means that you shouldn't get too excited just yet about the fact that
Guido van Rossum (36:00.400)
you figured out that we're made of atoms.
Guido van Rossum (36:02.240)
Right, and the same about taking facts as our sensory organs take them in and turning
Guido van Rossum (36:11.920)
that into reasons and actions, that sort of, there are a lot of abstractions that we haven't
Lex Fridman (36:19.820)
quite figured out how to deal with those.
Guido van Rossum (36:23.960)
I mean, sometimes, I don't know if I can go on a tangent or not, so if I take a simple
Guido van Rossum (36:37.440)
program that parses, say I have a compiler that parses a program, in a sense the input
Guido van Rossum (36:45.640)
routine of that compiler, of that parser, is a sensing organ, and it builds up a mighty
Guido van Rossum (36:55.640)
complicated internal representation of the program it just saw, it doesn't just have
Guido van Rossum (37:01.960)
a linear sequence of bytes representing the text of the program anymore, it has an abstract
Guido van Rossum (37:08.200)
syntax tree, and I don't know how many of your viewers or listeners are familiar with
Guido van Rossum (37:15.480)
compiler technology, but there's…
Lex Fridman (37:18.680)
Fewer and fewer these days, right?
Guido van Rossum (37:21.880)
That's also true, probably.
Guido van Rossum (37:24.920)
People want to take a shortcut, but there's sort of, this abstraction is a data structure
Guido van Rossum (37:30.360)
that the compiler then uses to produce outputs that is relevant, like a translation of that
Guido van Rossum (37:37.480)
program to machine code that can be executed by hardware, and then that data structure
Guido van Rossum (37:47.880)
gets thrown away.
Guido van Rossum (37:50.600)
When a fish or a fly sees, sort of gets visual impulses, I'm sure it also builds up some
Guido van Rossum (38:02.560)
data structure, and for the fly that may be very minimal, a fly may have only a few, I
Guido van Rossum (38:10.000)
mean, in the case of a fly's brain, I could imagine that there are few enough layers of
Guido van Rossum (38:17.680)
abstraction that it's not much more than when it's darker here than it is here, well
Guido van Rossum (38:24.040)
it can sense motion, because a fly sort of responds when you move your arm towards it,
Lex Fridman (38:29.880)
so clearly its visual processing is intelligent, well, not intelligent, but it has an abstraction
Guido van Rossum (38:39.240)
for motion, and we still have similar things in, but much more complicated in our brains,
Guido van Rossum (38:46.440)
I mean, otherwise you couldn't drive a car if you couldn't, if you didn't have an
Lex Fridman (38:50.400)
incredibly good abstraction for motion.
Guido van Rossum (38:53.480)
Yeah, in some sense, the same abstraction for motion is probably one of the primary
Guido van Rossum (38:59.160)
sources of our, of information for us, we just know what to do, I think we know what
Guido van Rossum (39:05.080)
to do with that, we've built up other abstractions on top.
Guido van Rossum (39:08.280)
We build much more complicated data structures based on that, and we build more persistent
Guido van Rossum (39:14.120)
data structures, sort of after some processing, some information sort of gets stored in our
Guido van Rossum (39:20.320)
memory pretty much permanently, and is available on recall, I mean, there are some things that
Guido van Rossum (39:27.240)
you sort of, you're conscious that you're remembering it, like, you give me your phone
Guido van Rossum (39:34.040)
number, I, well, at my age I have to write it down, but I could imagine, I could remember
Guido van Rossum (39:39.560)
those seven numbers, or ten digits, and reproduce them in a while, if I sort of repeat them
Lex Fridman (39:46.240)
to myself a few times, so that's a fairly conscious form of memorization.
Guido van Rossum (39:53.320)
On the other hand, how do I recognize your face, I have no idea.
Guido van Rossum (39:57.800)
My brain has a whole bunch of specialized hardware that knows how to recognize faces,
Guido van Rossum (40:04.080)
I don't know how much of that is sort of coded in our DNA, and how much of that is
Guido van Rossum (40:10.200)
trained over and over between the ages of zero and three, but somehow our brains know
Lex Fridman (40:17.960)
how to do lots of things like that, that are useful in our interactions with other humans,
Lex Fridman (40:26.000)
without really being conscious of how it's done anymore.
Guido van Rossum (40:29.880)
Right, so our actual day to day lives, we're operating at the very highest level of abstraction,
Lex Fridman (40:36.200)
we're just not even conscious of all the little details underlying it.
Guido van Rossum (40:39.760)
There's compilers on top of, it's like turtles on top of turtles, or turtles all the way
Guido van Rossum (40:43.360)
down, there's compilers all the way down, but that's essentially, you say that there's
Guido van Rossum (40:48.200)
no magic, that's what I, what I was trying to get at, I think, is with Descartes started
Guido van Rossum (40:54.920)
this whole train of saying that there's no magic, I mean, there's all this beforehand.
Guido van Rossum (40:59.600)
Well didn't Descartes also have the notion though that the soul and the body were fundamentally
Lex Fridman (41:06.120)
separate?
Guido van Rossum (41:07.120)
Separate, yeah, I think he had to write in God in there for political reasons, so I don't
Guido van Rossum (41:11.800)
know actually, I'm not a historian, but there's notions in there that all of reasoning, all
Guido van Rossum (41:17.880)
of human thought can be formalized.
Guido van Rossum (41:20.120)
I think that continued in the 20th century with Russell and with Gadot's incompleteness
Guido van Rossum (41:28.480)
theorem, this debate of what are the limits of the things that could be formalized, that's
Guido van Rossum (41:33.120)
where the Turing machine came along, and this exciting idea, I mean, underlying a lot of
Guido van Rossum (41:37.960)
computing that you can do quite a lot with a computer.
Guido van Rossum (41:43.160)
You can encode a lot of the stuff we're talking about in terms of recognizing faces and so
Guido van Rossum (41:47.640)
on, theoretically, in an algorithm that can then run on a computer.
Lex Fridman (41:53.960)
And in that context, I'd like to ask programming in a philosophical way, what does it mean
Lex Fridman (42:05.040)
to program a computer?
Lex Fridman (42:06.480)
So you said you write a Python program or compiled a C++ program that compiles to some
Guido van Rossum (42:13.360)
byte code, it's forming layers, you're programming a layer of abstraction that's higher, how
Lex Fridman (42:21.200)
do you see programming in that context?
Lex Fridman (42:24.920)
Can it keep getting higher and higher levels of abstraction?
Guido van Rossum (42:29.800)
I think at some point the higher levels of abstraction will not be called programming
Lex Fridman (42:35.960)
and they will not resemble what we call programming at the moment.
Guido van Rossum (42:44.720)
There will not be source code, I mean, there will still be source code sort of at a lower
Guido van Rossum (42:52.080)
level of the machine, just like there are still molecules and electrons and sort of
Guido van Rossum (42:59.320)
proteins in our brains, but, and so there's still programming and system administration
Lex Fridman (43:09.120)
and who knows what, to keep the machine running, but what the machine does is a different level
Guido van Rossum (43:15.960)
of abstraction in a sense, and as far as I understand the way that for the last decade
Guido van Rossum (43:23.060)
or more people have made progress with things like facial recognition or the self driving
Guido van Rossum (43:28.440)
cars is all by endless, endless amounts of training data where at least as a lay person,
Lex Fridman (43:38.200)
and I feel myself totally as a lay person in that field, it looks like the researchers
Guido van Rossum (43:47.420)
who publish the results don't necessarily know exactly how their algorithms work, and
Guido van Rossum (43:57.400)
I often get upset when I sort of read a sort of a fluff piece about Facebook in the newspaper
Guido van Rossum (44:04.840)
or social networks and they say, well, algorithms, and that's like a totally different interpretation
Guido van Rossum (44:12.680)
of the word algorithm, because for me, the way I was trained or what I learned when I
Guido van Rossum (44:19.240)
was eight or ten years old, an algorithm is a set of rules that you completely understand
Guido van Rossum (44:25.920)
that can be mathematically analyzed and you can prove things.
Guido van Rossum (44:30.720)
You can like prove that Aristotelian sieve produces all prime numbers and only prime
Guido van Rossum (44:37.840)
numbers.
Lex Fridman (44:38.840)
Yeah.
Lex Fridman (44:39.840)
So I don't know if you know who Andrej Karpathy is, I'm afraid not.
Lex Fridman (44:44.360)
So he's a head of AI at Tesla now, but he was at Stanford before and he has this cheeky
Guido van Rossum (44:51.980)
way of calling this concept software 2.0.
Lex Fridman (44:56.480)
So let me disentangle that for a second.
Lex Fridman (45:00.120)
So kind of what you're referring to is the traditional, the algorithm, the concept of
Guido van Rossum (45:06.080)
an algorithm, something that's there, it's clear, you can read it, you understand it,
Guido van Rossum (45:09.560)
you can prove it's functioning as kind of software 1.0.
Lex Fridman (45:14.800)
And what software 2.0 is, is exactly what you described, which is you have neural networks,
Guido van Rossum (45:21.920)
which is a type of machine learning that you feed a bunch of data and that neural network
Lex Fridman (45:26.600)
learns to do a function.
Guido van Rossum (45:30.200)
All you specify is the inputs and the outputs you want and you can't look inside.
Lex Fridman (45:35.220)
You can't analyze it.
Guido van Rossum (45:37.040)
All you can do is train this function to map the inputs to the outputs by giving a lot
Lex Fridman (45:41.920)
of data.
Lex Fridman (45:42.920)
And that's as programming becomes getting a lot of data.
Lex Fridman (45:47.040)
That's what programming is.
Guido van Rossum (45:48.920)
Well, that would be programming 2.0.
Lex Fridman (45:52.120)
To programming 2.0.
Guido van Rossum (45:53.800)
I wouldn't call that programming.
Lex Fridman (45:55.600)
It's just a different activity.
Guido van Rossum (45:57.480)
Just like building organs out of cells is not called chemistry.
Lex Fridman (46:02.640)
Well, so let's just step back and think sort of more generally, of course.
Lex Fridman (46:09.680)
But you know, it's like as a parent teaching your kids, things can be called programming.
Lex Fridman (46:18.080)
In that same sense, that's how programming is being used.
Guido van Rossum (46:22.720)
You're providing them data, examples, use cases.
Lex Fridman (46:27.080)
So imagine writing a function not by, not with for loops and clearly readable text,
Lex Fridman (46:36.680)
but more saying, well, here's a lot of examples of what this function should take.
Lex Fridman (46:42.760)
And here's a lot of examples of when it takes those functions, it should do this.
Lex Fridman (46:47.860)
And then figure out the rest.
Lex Fridman (46:50.280)
So that's the 2.0 concept.
Lex Fridman (46:52.640)
And so the question I have for you is like, it's a very fuzzy way.
Lex Fridman (46:58.560)
This is the reality of a lot of these pattern recognition systems and so on.
Guido van Rossum (47:01.680)
It's a fuzzy way of quote unquote programming.
Lex Fridman (47:05.400)
What do you think about this kind of world?
Lex Fridman (47:09.160)
Should it be called something totally different than programming?
Guido van Rossum (47:13.640)
If you're a software engineer, does that mean you're designing systems that are very, can
Guido van Rossum (47:21.000)
be systematically tested, evaluated, they have a very specific specification and then this
Lex Fridman (47:28.140)
other fuzzy software 2.0 world, machine learning world, that's something else totally?
Lex Fridman (47:33.520)
Or is there some intermixing that's possible?
Guido van Rossum (47:41.000)
Well the question is probably only being asked because we don't quite know what that software
Guido van Rossum (47:48.600)
2.0 actually is.
Lex Fridman (47:51.400)
And I think there is a truism that every task that AI has tackled in the past, at some point
Guido van Rossum (48:02.960)
we realized how it was done and then it was no longer considered part of artificial intelligence
Lex Fridman (48:09.160)
because it was no longer necessary to use that term.
Guido van Rossum (48:15.200)
It was just, oh now we know how to do this.
Lex Fridman (48:21.600)
And a new field of science or engineering has been developed and I don't know if sort
Guido van Rossum (48:30.320)
of every form of learning or sort of controlling computer systems should always be called programming.
Lex Fridman (48:39.000)
So I don't know, maybe I'm focused too much on the terminology.
Lex Fridman (48:43.720)
But I expect that there just will be different concepts where people with sort of different
Guido van Rossum (48:56.200)
education and a different model of what they're trying to do will develop those concepts.
Guido van Rossum (49:07.920)
I guess if you could comment on another way to put this concept is, I think the kind of
Guido van Rossum (49:17.240)
functions that neural networks provide is things as opposed to being able to upfront
Guido van Rossum (49:23.480)
prove that this should work for all cases you throw at it.
Lex Fridman (49:28.720)
All you're able, it's the worst case analysis versus average case analysis.
Guido van Rossum (49:32.320)
All you're able to say is it seems on everything we've tested to work 99.9% of the time, but
Lex Fridman (49:39.800)
we can't guarantee it and it fails in unexpected ways.
Guido van Rossum (49:44.160)
We can't even give you examples of how it fails in unexpected ways, but it's like really
Lex Fridman (49:48.080)
good most of the time.
Lex Fridman (49:50.120)
Is there no room for that in current ways we think about programming?
Guido van Rossum (50:00.720)
programming 1.0 is actually sort of getting to that point too, where the sort of the ideal
Guido van Rossum (50:11.080)
of a bug free program has been abandoned long ago by most software developers.
Lex Fridman (50:21.120)
We only care about bugs that manifest themselves often enough to be annoying.
Lex Fridman (50:30.120)
And we're willing to take the occasional crash or outage or incorrect result for granted
Guido van Rossum (50:40.680)
because we can't possibly, we don't have enough programmers to make all the code bug free
Lex Fridman (50:47.600)
and it would be an incredibly tedious business.
Lex Fridman (50:50.200)
And if you try to throw formal methods at it, it becomes even more tedious.
Lex Fridman (50:56.320)
So every once in a while the user clicks on a link and somehow they get an error and the
Lex Fridman (51:05.520)
average user doesn't panic.
Guido van Rossum (51:07.360)
They just click again and see if it works better the second time, which often magically
Lex Fridman (51:14.840)
it does, or they go up and they try some other way of performing their tasks.
Lex Fridman (51:21.600)
So that's sort of an end to end recovery mechanism and inside systems there is all
Guido van Rossum (51:29.880)
sorts of retries and timeouts and fallbacks and I imagine that that sort of biological
Guido van Rossum (51:39.120)
systems are even more full of that because otherwise they wouldn't survive.
Lex Fridman (51:46.320)
Do you think programming should be taught and thought of as exactly what you just said?
Guido van Rossum (51:54.160)
I come from this kind of, you're always denying that fact always.
Guido van Rossum (52:01.560)
In sort of basic programming education, the sort of the programs you're having students
Guido van Rossum (52:12.680)
write are so small and simple that if there is a bug you can always find it and fix it.
Guido van Rossum (52:23.480)
Because the sort of programming as it's being taught in some, even elementary, middle schools,
Guido van Rossum (52:29.720)
in high school, introduction to programming classes in college typically, it's programming
Lex Fridman (52:36.680)
in the small.
Guido van Rossum (52:38.920)
Very few classes sort of actually teach software engineering, building large systems.
Lex Fridman (52:47.560)
Every summer here at Dropbox we have a large number of interns.
Guido van Rossum (52:51.360)
Every tech company on the West Coast has the same thing.
Guido van Rossum (52:56.720)
These interns are always amazed because this is the first time in their life that they
Guido van Rossum (53:02.520)
see what goes on in a really large software development environment.
Guido van Rossum (53:12.920)
Everything they've learned in college was almost always about a much smaller scale and
Guido van Rossum (53:20.280)
somehow that difference in scale makes a qualitative difference in how you do things and how you
Lex Fridman (53:27.840)
think about it.
Guido van Rossum (53:29.600)
If you then take a few steps back into decades, 70s and 80s, when you were first thinking
Guido van Rossum (53:36.300)
about Python or just that world of programming languages, did you ever think that there would
Lex Fridman (53:41.840)
be systems as large as underlying Google, Facebook, and Dropbox?
Lex Fridman (53:46.720)
Did you, when you were thinking about Python?
Guido van Rossum (53:51.440)
I was actually always caught by surprise by sort of this, yeah, pretty much every stage
Lex Fridman (53:57.520)
of computing.
Lex Fridman (53:59.680)
So maybe just because you've spoken in other interviews, but I think the evolution of programming
Guido van Rossum (54:07.280)
languages are fascinating and it's especially because it leads from my perspective towards
Guido van Rossum (54:13.080)
greater and greater degrees of intelligence.
Guido van Rossum (54:15.640)
I learned the first programming language I played with in Russia was with the Turtle
Guido van Rossum (54:21.880)
logo.
Lex Fridman (54:22.880)
Logo, yeah.
Lex Fridman (54:24.840)
And if you look, I just have a list of programming languages, all of which I've now played with
Lex Fridman (54:29.960)
a little bit.
Guido van Rossum (54:30.960)
I mean, they're all beautiful in different ways from Fortran, Cobalt, Lisp, Algol 60,
Guido van Rossum (54:36.640)
Basic, Logo again, C, as a few, the object oriented came along in the 60s, Simula, Pascal,
Guido van Rossum (54:46.160)
Smalltalk.
Lex Fridman (54:47.560)
All of that leads.
Guido van Rossum (54:48.560)
They're all the classics.
Lex Fridman (54:49.560)
The classics.
Guido van Rossum (54:50.560)
Yeah.
Lex Fridman (54:51.560)
The classic hits, right?
Guido van Rossum (54:52.560)
Steam, that's built on top of Lisp.
Guido van Rossum (54:58.280)
On the database side, SQL, C++, and all of that leads up to Python, Pascal too, and that's
Guido van Rossum (55:05.900)
before Python, MATLAB, these kind of different communities, different languages.
Lex Fridman (55:10.960)
So can you talk about that world?
Guido van Rossum (55:13.240)
I know that sort of Python came out of ABC, which I actually never knew that language.
Guido van Rossum (55:18.680)
I just, having researched this conversation, went back to ABC and it looks remarkably,
Guido van Rossum (55:24.400)
it has a lot of annoying qualities, but underneath those, like all caps and so on, but underneath
Lex Fridman (55:31.240)
that, there's elements of Python that are quite, they're already there.
Guido van Rossum (55:35.720)
That's where I got all the good stuff.
Lex Fridman (55:37.540)
All the good stuff.
Guido van Rossum (55:38.540)
So, but in that world, you're swimming these programming languages, were you focused on
Guido van Rossum (55:41.580)
just the good stuff in your specific circle, or did you have a sense of what is everyone
Lex Fridman (55:48.080)
chasing?
Lex Fridman (55:49.080)
You said that every programming language is built to scratch an itch.
Lex Fridman (55:57.000)
Were you aware of all the itches in the community?
Lex Fridman (55:59.920)
And if not, or if yes, I mean, what itch were you trying to scratch with Python?
Guido van Rossum (56:05.080)
Well, I'm glad I wasn't aware of all the itches because I would probably not have been able
Lex Fridman (56:12.040)
to do anything.
Guido van Rossum (56:14.040)
I mean, if you're trying to solve every problem at once, you'll solve nothing.
Lex Fridman (56:19.760)
Well, yeah, it's too overwhelming.
Lex Fridman (56:23.880)
And so I had a very, very focused problem.
Guido van Rossum (56:28.360)
I wanted a programming language that sat somewhere in between shell scripting and C. And now,
Guido van Rossum (56:41.880)
arguably, there is like, one is higher level, one is lower level.
Lex Fridman (56:48.720)
And Python is sort of a language of an intermediate level, although it's still pretty much at
Guido van Rossum (56:56.760)
the high level end.
Guido van Rossum (57:00.560)
I was thinking about much more about, I want a tool that I can use to be more productive
Guido van Rossum (57:11.200)
as a programmer in a very specific environment.
Lex Fridman (57:16.640)
And I also had given myself a time budget for the development of the tool.
Lex Fridman (57:22.280)
And that was sort of about three months for both the design, like thinking through what
Guido van Rossum (57:29.340)
are all the features of the language syntactically and semantically, and how do I implement the
Guido van Rossum (57:38.900)
whole pipeline from parsing the source code to executing it.
Lex Fridman (57:43.680)
So I think both with the timeline and the goals, it seems like productivity was at the
Guido van Rossum (57:51.440)
core of it as a goal.
Lex Fridman (57:54.040)
So like, for me in the 90s, and the first decade of the 21st century, I was always doing
Guido van Rossum (58:01.280)
machine learning, AI programming for my research was always in C++.
Lex Fridman (58:07.620)
And then the other people who are a little more mechanical engineering, electrical engineering,
Guido van Rossum (58:14.240)
are MATLABby.
Lex Fridman (58:15.240)
They're a little bit more MATLAB focused.
Guido van Rossum (58:18.520)
Those are the world, and maybe a little bit Java too.
Lex Fridman (58:21.200)
But people who are more interested in emphasizing the object oriented nature of things.
Lex Fridman (58:29.160)
So within the last 10 years or so, especially with the oncoming of neural networks and these
Guido van Rossum (58:34.920)
packages that are built on Python to interface with neural networks, I switched to Python
Lex Fridman (58:41.360)
and it's just, I've noticed a significant boost that I can't exactly, because I don't
Guido van Rossum (58:47.120)
think about it, but I can't exactly put into words why I'm just much, much more productive.
Guido van Rossum (58:52.840)
Just being able to get the job done much, much faster.
Lex Fridman (58:56.400)
So how do you think, whatever that qualitative difference is, I don't know if it's quantitative,
Guido van Rossum (59:01.880)
it could be just a feeling, I don't know if I'm actually more productive, but how
Lex Fridman (59:07.280)
do you think about...
Guido van Rossum (59:08.280)
You probably are.
Lex Fridman (59:09.280)
Yeah.
Guido van Rossum (59:10.280)
Well, that's right.
Guido van Rossum (59:11.880)
I think there's elements, let me just speak to one aspect that I think that was affecting
Guido van Rossum (59:15.400)
my productivity is C++ was, I really enjoyed creating performant code and creating a beautiful
Guido van Rossum (59:26.160)
structure where everything that, you know, this kind of going into this, especially with
Guido van Rossum (59:31.000)
the newer and newer standards of templated programming of just really creating this beautiful
Guido van Rossum (59:37.080)
formal structure that I found myself spending most of my time doing that as opposed to getting
Guido van Rossum (59:42.000)
it, parsing a file and extracting a few keywords or whatever the task was trying to do.
Lex Fridman (59:47.520)
So what is it about Python?
Lex Fridman (59:49.980)
How do you think of productivity in general as you were designing it now, sort of through
Lex Fridman (59:54.520)
the decades, last three decades, what do you think it means to be a productive programmer?
🔗 相关节目