YouSaid · the spoken record

Guido van Rossum

lines on the record: 185
first: 2022-11-26
most recent: 2022-11-26
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“It is fascinating to see what abstraction levels are built on top of individual humans. And how you have whole societies. That sort of have a similar self preservation I didn't know what it is, instinct, nature, abstraction as the individuals have and the cells have.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Isn't it interesting that you sort of see the same thing appearing at different levels though because you have like Cells that create new cells And eventually, that builds a whole organism, but then the animal or the plant or the human. Has its own mechanism of replication that is sort of connected in a very complicated way to the mechanism of replication of the cells. And then if you look inside the cell, if you see how DNA and proteins are connected, then there is yet another completely different mechanism whereby proteins are mass produced. Using enzymes and a little bit of code from DNA. And of course, viruses break into it at that level.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Parallel circuitry in our brains, and eventually we use that to simulate a single threaded much, much higher level interpreter.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Maybe, maybe not. Yeah, I did. I mean, it's possible. I mean, if you think of human consciousness, if that's even the right concept. It's interesting that sort of we have this super parallel brain that does all these incredible parallel operations like image recognition. I recognize your face. Huge amount of processing that goes on in parallel. There's lots of nerves between my eyes and my brain. The brain does a whole bunch of stuff all at once because it's actually really slow circuits, but there are many of them that all work together. On the other hand, when I'm speaking, everything is completely sequential I have to sort of string words together one at a time. And when I'm thinking about stuff, when I'm understanding the world. I'm also thinking of everything like one step at a time. And so we've got all this incredible.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, I feel I sort of covered this middle layer of the technology stack where it sort of peters out blow the... Level of NAND And at the top, I sort of lose track when it gets to machine learning.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“The other day as a sort of Mental exercise I was trying to figure out if I could build a Flip-flop circuit out of relays. Was just sort Trying to remember, oh, how does a relay work? Yeah, there's like this electromagnetic force that pulls a switch open or shut. And you can have like it can open one switch and another, shut another. You can have multiple contacts that go at once. And how many relays do I really need to sort of represent one bit of information? Can the relay just feed on itself? I don't think I got to the final solution, but it was fun that I. I could still do a little bit of problem solving and thinking at that level.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“I started building little digital circuits out of Nandgates that I built myself with transistors and resistors. So I sort of, I feel very blessed that with that start when I was a teenager. I learned some of the basic, at least concepts that go into building a computer. And I sort of every part I have some understanding what. What it's for and why it's dare and how it works. And I can forget about all that most of the time, but I sort of, I enjoy knowing, oh, if you go deeper, at some point you get to Nand Gates and have others and shift registers. It comes to the point of how do you actually make a chip out of silicon, I have no idea. That's just magic to me.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“It'll eventually become sort of a legacy language that plays an important role, but that most people have never heard of and don't need to know about. Just like all kinds of... Basic structures in biology, like mitochondria.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Our programmers' jobs threatened by the existence of stack overflow? I don't think so. It helps you take care of the boring stuff. And you shouldn't try to use it to do something that you have no way of understanding what you're doing yet. A tool like that is always best when the question you're asking is, please remind me of how I do this, which I... I could look up how to do it, but right now I've forgotten whether the method is called foo or bar or how you, what the shape of the API is, does it use a builder object or a constructor or a factory? Something else, and what are the parameters, it serves that role. It's like a great assistant. But the creative work of sort of deciding what you want the code to do is totally yours.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“It doesn't save me a lot of thinking, but since I'm a poor typist, I'm very much appreciative of all the typing it does for me much better actually than the previous generation of suggestions that are also still built in VS Code. Where when you hit like a dot, it tries to guess what the type is of the variable to the left of the dot, and then it gives you a list, a pop-down menu of what the attributes of that object are. But copilot is much, much smoother than that.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, it writes a lot of code for me. And usually it's slightly wrong, but it still saves me typing. Because all I have to do is change one word in a line of text that otherwise it generated perfectly. And like, how many times are you looking for like, oh, what was I doing this morning? I was looking for a begin marker and I was looking for an end marker. And so begin is blah, blah, blah search for begin. This is the begin token. And then the next line, I type E and it completes the whole line with end instead of begin. That's a very simple example. Sometimes it sort of, if I name my function right, it writes a five or ten line function.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, you have to practice. You can memorize the syntax. But, well, I couldn't, but maybe someone else can, but that doesn't make you a coder.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, you can't expect to learn Python from a one hour video. Carvalho, blanking out on the name of someone who... Who wrote a very funny blog post where he said, I see all these ads for things like learn Python in ten days or so. And he said the goal should be learn Python in 10 years.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Find something you actually want to do with it. If you say, I want to learn skill X, That's not enough motivation. You need to pick something And it can be a crazy problem you want to solve. It can be completely unrealistic. But something that challenges you into actually learning In some language.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“And at least when I was there, he had not checked out or anything. He was good CEO, but he had started out as the technical inventor or co-inventor. And so he was someone who... I don't know if he always aspired that, I think when he was 16, he already started a company. So maybe he did, but he sort of. Turned out that he did have the personal sort of skill set needed to grow and stay on top. And other people sort of Brilliant engineers and horrible at management. I count myself, at least in the second category.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“If that's your aspiration, what can also happen is that Someone is brilliant engineer and sort of builds a great first version of a product. And has no aspirations to then become a manager and grow the team from five people to twenty people to 100 people to 1,000 people. Be in charge of hiring and meetings, and they move on to inventing another crazy thing inside the same company or sometimes they They found a startup, or they moved to a different, great larger, small company. There's all sorts of models. Sometimes people sort of do have this whole trajectory from engineer buckling down writing code. Not nine to five, but more like noon till midnight, seven days a week. And coming up with a product and sort of Staying in charge. I mean, if you take Drew Houston, Dropbox's founder, he is still the CO.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“I've never met him, but I hear He's just a really sharp thinker But he also has an incredible business sense. He took the organization that had very solid pieces but that was also. Struggling With all sorts of shameful things, especially the Steve Ballmer time, I imagine in part through his personal charm and thinking and of course the great trust that the rest of the leadership has in him, he managed to really turn the company around and sort of Change it from openly hostile to open source to actively embracing open source. And that doesn't mean that suddenly Excel is going to go open source. But that means that there's room for a product like VS Code, which is open source.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Just have to know what they are. That's No logic or reason to the assignment of the keyboard shortcuts because they go back even longer than 35 years.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Lot of the engineering sort of, if you look deep inside Excel, Some very good engineering, very, very impressive stuff. Okay, now I need to definitely learn”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Microsoft is a very interesting company because it has this incredible, very long and varied history and this amazing catalog of products that many of which also date way back. I mean I've been talking to a bunch of Excel people lately and Excel is like 35 years old. And they can still read spreadsheets that they might find on an old floppy drive.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“That's like why I joined Microsoft. I mean, after retiring and thinking that I would stay retired for the rest of my life, which of course was a ridiculous thought. I was done working for a bit and then the pandemic made me realize that work can also provide a source of fulfillment.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Lot of internal sort of hand wringing about things like that. But the product itself always worked very smoothly and still does.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Had the tendency of sort of making a big plan, putting the whole company behind that plan for a year, and then evaluate and then suddenly find that. Everything was wrong about the plan, and then they had to do something completely different. So there was the annual engineering reorg was sort of an unpleasant tradition at Dropbox because like, oh, there's a new VP of engineering. And so now all the directors are being reshuffled. And this guy was in charge of infrastructure one year and the next year he was made in charge of, I don't know, product development.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah. Oh, yeah, there's a bug in Meldu processing. Let someone else decide whether that's important to solve or not. I'll stick to typing in the async I.O. and the faster interpreter”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“It was a source of stress for me personally, but it was beneficial for the community because people sort of over time had learned how I was thinking and could predict. But howard would decide about a particular issue and not always perfectly, of course, but there was like Wasn't a lot of jerking around like this year. We're all this year, the Democrats are in power and we're doing these kind of things and now the Republicans are in power and they roll all that back and do those kind of things. There is a clear fairly straight path ahead. And so fortunately the successor structure with the steering council has sort of found a similar way of leading the community in a fairly steady direction without stagnating. And for me personally, it's more fun because there are things I can just ignore.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, the benefits for the community would be things like. Clarity of vision and sort of Clear direction because I had certain ideas in mind when I created Python. And while I sort of let myself be influenced by many other ideas as Python evolved and became More successful and more complex and more used. I also stuck to certain principles and it's still hard to say what are Python's core principles. But the fact that I was playing that role and sort of always very active grew the community in a certain way. Modeled to the community how to think about How to solve a certain problem? Well,”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“I probably held on to the position too long. I remember being just Extremely stressed for a long time. It wasn't very clear to me. What was leading? What was causing the stress? And looking back. I should have sort of relinquished my central role as BDFL sooner.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Tend to like Python's approach to open source in particular because it's sort of It's almost egalitarian. There's a little hierarchy. There's obviously some because you all need to decide whether you drive on the left or the right side of the road sometimes. But there is a lot of access for people with little power. You don't have to work for a big tech company to make a difference in the Python world. We have affordable events that really care about community and support people and sort of the community is like a big deal at our conferences and in the PSF. When the PSF funds events, it's always about growing the community. The PSF funds very little development that they do some, but most of the develops out. Is to community fostering things.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Was very expensive product. And so universities especially disliked it because it was a price per seat, I remember hearing.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“For arrays of numbers yet, whereas in Python you have it. And so more and more scientists at different places doing different work discovered Python and then people who had an idea for an important new fundamental library decided, oh, Python is actually already known to our users. So let's use Python as the user interface. I think that's how tensor, I imagine at least that's how TensorFlow ended up with Python as the user interface.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“And at various points, small improvements were made and more people got in touch with using Python to derive these libraries. Of interesting algorithms. And once you have a bunch of scientists who are working on similar problems, say they're all working on stuff that comes in from the Hubble Space Telescope, but they're looking at different things. Some are looking at stars in this galaxy. Others are looking at galaxies. The math is completely different. The underlying libraries are still the same. And so they exchange. They say, well, I wrote this Python program, or I wrote a Python library to solve this class of problems. And the other guys either say, oh, I can use let library too, or if you make a few changes, I can use that library too. Why start from scratch in Perl or JavaScript, where there's not that infrastructure?”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Gradually, some libraries started appearing that did very fundamental stuff with arrays of numbers in Python. I mean, when I first created Python, I was And everything was like objects and strings and like Python was good and fast at string manipulation in objects obviously. But arrays of numbers were not very efficient and the multidimensional arrays didn't even exist in the language at all. But there were people who realized that Python had extensibility Was flexible enough that they could write third party packages that did support large arrays of numbers and operations on them very efficiently. And somehow they got a foothold through sort of different parts of the scientific community. I remembered that the Hubble Space Telescope people in Baltimore were somehow big Python fans in the late 90s.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“And those libraries still work, but the scientists that use the libraries use them to solve continuously different specific applications and answer different questions. And so those poor scientists were required to use, say, Fortran because Fortran was the language that the library was written in. And then the scientist would have to write an application that sort of uses the library to solve a particular equation or set of answer a set of questions. And the same for C++. And there's interoperability. So the dusty decks are written either in C or Fortran. And so Paul Dubois was one of the people who I think in the mid 90s saw that you needed a higher level language for the scientists. To sort of tie together the fundamental mathematical algorithms of linear algebra and other stuff. And so”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“The reasons why Python became the lingua franca of scientific code and Machine learning in particular and data science. It really had a lot to do with. Anything was better than C or C. Recently, a guy who worked at Lawrence Livermore National Laboratories in the sort of computing division. Wrote me his memoirs and he had his own view of how he helped something he called computational steering into existence. And this was the idea that you take libraries that in his days were written in Fortran that solved universal mathematical problems.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, part of it is an effect that's as simple as we're all driving on the right side of the road, right? It's compatibility And part of it is. Not quite as fundamental as driving on the right side of the road, which you have to do for safety reasons. I mean, you have to agree on something. They could have picked JavaScript or Perl. There was a time in the early 2000s that it really looked like Perl was going to dominate like biosciences. Because DNA search was all based on regular expressions, and Pearl has the fastest and most comprehensive regular expression engine still does.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Would be better off not existing at all at this point because there might be a better third party alternative anyway and we're sort of slowly removing those. Often those are things that I sort of I spiked somewhere in 1992 or 1993. And if you look through the commit history, it's very sad. All cosmetic changes like changes in the indentation style or the name of this other standard library module got changed. Nothing of any substance. The API is identical to what it was 20 years ago.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, usually when something's missing from the standard library nowadays, It is a relatively new idea, and there is a third party implementation or maybe possibly multiple third party implementations, but they evolve at a much higher rate than they could when they're in the standard library. So it would be a big reduction in activity to incorporate things like that in the standard library. So I like that there is a lively package ecosystem and that sort of recent trends in the standard library are actually that we're doing the occasional sprint cleaning where we're just Choosing some modules that have not had a lot of change in a long time and that maybe”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Oh my, that's Should really have a standard answer for that question, but like a positive standard answer, but my current standard answer is that I'm not a big user of third-party packages. When I write Python code, I'm usually developing some tooling around building Python itself. And the last thing we want is dependencies on third party packages. So I tend to just use the standard library.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Test the waters, see what they need to syntactically to be able to compile at all. Maybe they're using. Functions that are defined by the Python 3 runtime that won't be in the Python 4 runtime, those functions will not work. They'll have to find an alternative. But they can experiment with that and sort of write test applications. And that would be a way to transition. And that could be a series of releases where the Python 4 is more and more imminent. We have supported more and more third-party extension modules to have solid support that works for NOL Python, for that new API. And then sort of Python 4.0 is like the official moment that the mayor comes out and cuts the ribbon and now Python Now, the sort of no gill mode is the default and maybe the only mode there is.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“We probably have to do something like Several years before, who knows, maybe five years earlier, like 3.15, we would have to say, and I'm just making the specific numbers up, but at some point we'd have to say, The nogill Python could be an option. It might be a compile time option. If you want to use Nogil Python, you have to recompile Python from source for your platform using your tool set. All you have to do is change one configuration variable and then you just run make or configure and make and it will build it for you. But now you also have to use the no-gill compatible versions of all extension modules you want to use. And so as long as many extension modules don't have fully functional Sort of variants that work in the no-gill world that's not a very practical thing for Python users, but it would allow extension developers.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Things that are frequently accessed by C extensions will be different. And so for a pure Python user, 4.0 would be a breeze, except that there are very few pure Python users left because everybody who is using Python for something significant is using third-party extensions. They're like, I don't know several hundreds of thousands of third-party extensions on the PyPI service. And I'm not saying they're all good, but there is a large list of extensions that would have to do work. And some of those extensions are currently already low on maintainers. And they're struggling to keep afloat.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Suppose that's the time when we flip the switch to 4.0 will not have a gill. Imagine it was like that. I would probably say That particular year, the release that we name 4.0 will be syntactically, it will not have any new syntactical features. No new modules in the standard library, no new built-in functions. Everything will be at the Python level, will be purely compatible with Python 3.19. However Extension modules will have to make a change. They all have to be recompiled. They will not have the same Binary interface The semantics and APIs for some.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, so Here is a concrete thought I've had, and I'm not unique, but not everyone agrees with this. So this is definitely a personal opinion. If we were to try something like that no guild Python, My expectation is that. It would feel just different enough. At least for the part of the Python ecosystem that is heavily based on C extensions. And that is like the entire machine learning data science scientific Python world is all based on C extensions for Python. And so Those people would likely feel the pain the most because they Even if we don't change anything about the syntax of the language and the semantics of the language when you're writing Python code, we could even say suppose that after Python, say 319 instead of 320, we'll have 4.0.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, the conference was going to be, I think, late April that year. So that was a very difficult decision to cancel it. Did anyway, if we're going to have a Python 4, we're going to have to have both a different reason for having that and a different process for managing the transition.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“I'm not a very poetic person, and I don't shed tears like that, but no. We actually had planned a party, but the party was planned for the US Python conference that year, which never happened, of course, because of the pandemic.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“It was purely a little symbolic moment. To signal to the remaining users that There was no longer going to be any new releases or support for Python 2.7.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, well, we're. As a community and as a core development team, we have a large amount of painful memories about the Python 3 transition. Which is one reason that sort of everybody is happy that we've decided there's not going to be a 4.0 at least not anytime soon. And if there is going to be one. Because clearly we underestimated the pain the transition caused for our users in the Python 3 case. Had we known we could have sort of designed Python 3 somewhat differently without making it any worse? We just thought that we had a good plan, but we underestimated what sort of the users were capable of when it comes to that kind of transition.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“There are enough people needing the speed of multiple threads with their Python programs that it's worth to sort of take that performance hit and that complexity hit. And I feel that the GIL actually is a pretty nice Goldilocks point between no threads and all threads all the time. But not everybody agrees on that. So that is definitely a possible future. The subinterpreters look like a fairly safe bet for 312, so say a year from now.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“The interesting thing is that there is still a possible future where we are actually going to or where we could experiment at least with that because there is a guy working for Facebook who has developed a fork of CPython that he called the no-gill interpreter where he removed the gill and made a whole bunch of optimizations so that the single threaded case doesn't run too much slower. And multi-threaded case will actually use all the cores that you have. And so that would be an interesting possibility if we would be willing as Python core developers to actually Maintain that code indefinitely. And if we're willing to put up with the additional complexity of the interpreter and the additional sort of overhead for the single threaded case, and I'm personally not convinced that”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source
“It's hidden from you, but you have to spend more time communicating between those programs because the sort of attractive thing about the multi-threaded model is that the threads can share objects. At the same time, that's also the downfall of the multi-threaded programming model. Because when you do share objects, and you didn't necessarily intend to share them or there were aspects of those objects that were not reusable, you get all kinds of concurrency bugs. The reason I wrote that little blog post about semaphores was that concurrency bugs are just harder. It would be nice if Python had no global interpreter lock and it had the so-called free threading. But it would also cause a lot more software bugs.”
2022-11-26 · Lex Fridman Podcast · #341 – Guido van Rossum: Python and the Future of Programming · IDENTIFIED FROM THE TRANSCRIPT · source