Full podcast episode: th-cam.com/video/-DVyjdw4t9I/w-d-xo.html Lex Fridman podcast channel: th-cam.com/users/lexfridman Guest bio: Guido van Rossum is the creator of Python programming language.
Just graduated with a degree in software engineering. We still had to learn how to make a basic processor and made a pong game from just binary inputs. The fundamentals are still being taught. They’re just being forgotten quickly with trying to learn all the new frameworks and dev tools out there.
Which, from an educational standpoint, is fair enough. People do tend to forget what they either don't use or don't find interesting, though the knowledge might still be usefull for someone.
I think there’s a middle ground. I appreciated the low level concepts I learned in UC schooling. It’s helped. I’m also pretty sure I didn’t have to program battleship in assembly (for two players where it’s human v human or human v computer) to really understand low level programming. Could have done without that suffering.
Congrats on the degree. The problem with prioritizing learning these lower-level parts is it holds back and gatekeeps technological advancement. We need people that can do the job, not necessarily build the tools to do the job. Those are specializations and separate jobs now. People train in these specializations but the more foundational people wont hire them and stuff stagnates.
@@colinmaharaj50 well C will live long past python so that won’t be a problem. I struggle to imagine a future of computing that isn’t extremely dependent on C.
You can look up the GNU Project's page on GCC. It has a steering committee that makes decisions for the project. And the organization behind GCC -- the GNU Project -- is the same as it's always been.
@@colinmaharaj50 Various other implementations of Python exist. My calculator has some non-CPython running on it, and I know off the top of my head that Jython and IronPython (C#) exist
@@samdcbu well we should consider it C is quite hard to develop in at times it could be nice if we ever do to C as we sort of did with assembly and make a language that compiles extremely efficiently to C but with higher capabilities and no python doesn’t really count for that
The thing that actually feels like that - in an unpleasant way, I might add - is Kotlin. It feels like speaking English except with an unusual form of Tourette's syndrome where you say some other punctuation like periods and brackets in the middle. It's actually disorienting! Python on the other hand is friendly without being patronizing
@@achrafaguenaou7825it’s kind of obvious that he was saying that it’s easier to get a job as a front end dev compared to the difficulty of getting any job with python. Backend, data science, whatever it may be- he’s saying it’s harder to get a job in those fields without a degree than it is to get a job as a front end dev. I felt like what he said was relatively straightforward so I don’t know how else to boil this one down.
1:40 Yeah, same here, I started with transistors and hand made digital circuits in 1991. My first digital circuit was a ramp generator to make a Digital multimeter. I did 8085 ASM, then I got a 286 when I was employed in my 1st (or 2nd) year. In office they had a 8086 IBM XT so I was more highly advanced at home. LOL.
At some point it will be hard to find a compromise between adding more features to the language, maintaining backwards compatibility and preserving simplicity, and being simple is one of the core design points of Python. I believe that is one of the reasons Python 2 and 3 are not compatible, they decided it was a more acceptable trade-off. But it can't continue forever, it will either stop or Python become new C++. Of course there will be new languages inspired by Python.
I actually don't think he's right here, and I think people who think this endless abstraction can continue are the same people who think that the world financial system can't collapse. We keep just stacking more shit, more layers of abstraction, but if we do that for a long enough time, then we're eventually going to have systems where not only does no one know actually how they work (which we have now), but we'll have entire systems where it will take multiple life times of research just to figure out how they work. Machine learning isn't the answer, because that's just adding more layers of abstraction--and obfuscated abstraction at that--which will hide how things work. Even worse, if it really can program beyond human capability, then we're not going to be able to understand the code it output, and as a result how can we audit it? We'll have to just "trust" the ai, which I think is not a good idea. I really feel like we're building the tower of babel, and we're headed for a very very big fall here.
This is probably why colleges start from the basics of computer science wrt computer architecture, digital logic, operating systems... Probably down the line their curriculum could also become abstracted.. Which is a terrifying thought to me.
The more complex the system the more complex the abstractions. Complexity causes more abstraction. If complexity of what software should do increases while no more abstraction layers are added it will be harder, not easier to understand how everything works compared to the situation where the rise in abstraction follows the rise in complexity. Your comment reminded me of a guy who claimed the busses cause rush hour. During the rush hour when people are getting back from work there are more buses on the street. When rush hour is over the number of buses on the street decreases. He concluded that buses are the cause of traffic. In reality, buses help make the traffic less bad and if the rise in traffic isn't followed by the rise in buses on the street, the traffic would be much worse.
Oh, I have just noticed that OP is the idiot who claimed C is not pass by value on a different video. Yeah, this is the kind of "expert" that we are dealing with. He actually isn't any better than that guy who I met on the bus who argued that the traffic is so bad because there are too many buses on the streets. Read my previous comment for context
I’m a data analyst at a big streaming company. We need faster languages than Python, SQL and R as datasets get ever larger. I work with pretty large datasets and building intermediary tables, optimising and sampling can only take you so far. These languages are getting inadequate for modern computing needs - especially with regards to data. Interested to see what new language supersedes them…
Even native Python will be 5 times as fast in 5 years. You'll probably have better C/C++ bindings, faster algorithms written into libraries, better multithreading support - and faster computers as well on the hardware side of things. Ultimately library support matters and that's where Python will stay unbeatable
On replication, fascinating thoughts. The interesting thing is that we are not generally aware of the layers upon layers of things that must fall in place for nature to create. And it's a miracle that it works so flawlessly almost everytime.
Obviously eventually it will be a legacy language, but the thing is, what will replace it and be easier to use? Python is one of the easiest and most supported languages to use right now, I don't see any other language replacing it for what it does unless you need some extremely efficient and fast code or a front-end web app.
50 - 100 years is an extremely long time in software. Consider how prevalent mobile applications have become in just 15 years. I agree that Python is currently in a great spot, but long-term, there's no telling what the future of software holds.
The good news is that the only way for it to be a legacy language is if a new and easier language comes around. And when that happens, it'd be easy for current python users to move along. Just like how easy it is for C developers to pick up python.
He makes a point clearly..might be a legacy one but can not be replaced.also it will take at least 50 + years for the conversion..and also will be so only when bots chooses the tech stack
Musicians with modular synths play with basic (CMOS 4xxx) logic daily... 🤷♂️ It is still fully useable, specially in rhythmical pattens 8 or 16 note patterns are very common all the time.
lol, software made on top of something unkown (like a programmer that does not how a cpu works), and those who make the cpu and know how it works, don't know in the end what electricity is in the end. Amazing
@@gonkong5638 won't be that long. Biggest deciding variable is development and widespread adoption of open source alternatives to the Linux kernel. That is unlikely to ever be needed, so C is safe, but C++ is more of an open question. Either Rust overtakes C++ within maybe a generation, or something shinier comes along cutting Rust down before it has a critical mass of projects. Unlike C, C++ has no real central nucleus of development though, so popularity is more variable with support for frameworks and language bindings.
Damn so there’s gonna be a new one in the future that will lead the job market? Definition of a legacy language: A ProgrammingLanguage which is still used in production code, but only in LegacySystems. In other words, where the language is typcally only selected when its use is a bona fide requirement, usually due to the need to interface with zillions of lines of code that already exist.
I think python is worth the learn. Let say Python is to slow for your use case so you make the real computation using Cpp or Rust, you can use python to call those package
I'm not invested in cybersecurity but I am a software engineer - Python will be around for a while, long enough at least for you to get familiar with coding and syntax in general. Once you're more well versed in writing programs for your use case then it will be much easier to switch to new languages need if needed and tackle more tough solutions.
Python for me is just to slow. That said I understand that it's taught in schools and is populair now. But vb5 was super populair when I was a kid/ teenager. These things often change over time quite quickly
Python is only slow if you're using it for the wrong purposes. The speed critical stuff should be compiled C/C++ code that can be imported into Python... which could then act as a general interface to it. This is the reason why Python is doing well in the SciComp community.
@@nasimicin Look at "Go" ;-) It's a little bit simplistic so some people who want the complexity of C++/Rust are disappointed. But it's quite simple and it compiles to machine code like C/C++/Pascal/COBOL/Fortran Another example would be "Julia" - a "faster Python", basically.
not sure , nim is good but we need a scripting language for some web/data science tools , i don't know why we still making more general use programming languages , C/C++ , go , python , js and rust could be the only languages that we use and other people just commit to them instead of creating new ones i can even delete go from the list , but they build a huge ecosystem around it (docker , kubernetes...) don't get me wrong i like trying languages , but a lof of time i don't really see the need tried D,V,Nim, julia ,peregrine , but i did not see a lot of added value
Of course, I am not a futurologist but I believe it is hard to find something more inspiring than C++ running on Linux. All other languages are sort of derivatives or shortcuts of the most probably powerful language human being pioneered. Sorry, this is only my humble point of view.
But C++ is just "C with fancy classes". And C is just "B but with less limitations". There was D as a "modern C++" but it failed by being just a little bit better so there is no real reason to use it. Maybe Rust would be different enough to become that "next C++"? ;-)
@@igorthelight You are right. I think however that Carbon will extend C++ abilities since it is back compatibility with C++. Anyway, I do not see the reason why people have to resign from using C++. The beauty is the fact that in CS everyone can find something which is the most suitable for solving a particular problem. I do not have anything against any language and I would like to avoid saying one language is better than the other. Good luck.
Assembly is faster but yes C++ strikes the best balance. Different languages serve different purposes. C, C++ are still better when speed of execution is desirable. Python is better from the view that you don't even need to be a professional programmer to write code in it and so a vast array of libraries gets built up
I think it's pretty naive to compare the human brain to a computer. Throughout history, humans have always compared the function of the brain to whatever the most advanced technology of the time was. People used to say the brain was like a clock or a watch. Now they think of it in terms of computers and code. The fact is, we know very little about how the chemistry of the brain translates to actual thought and perception, and comparisons to computers are just speculation based on existing technology we understand. Brain function, while we know some things, is still largely undiscovered territory. We don't know if machine learning / deep learning are really similar at all to brain function at all because we simply do not understand brain function. Machine learning can give the appearance of thought in some cases, but it doesn't really act like a human at all, and for the most part, it is very good a some things humans are bad at and very bad at some things humans are very good at. It's not even clear to me that humankind has the capacity to comprehend the mechanism of our own thought. Maybe we do, but we aren't anywhere close to that point yet.
Sadly python is flawed and not easy at all to understand for children. Try adding 0.1 three times like this: x = 0.1 + 0.1 + 0.1 THEN print( x ) gives 0.30000000000000004 , yayy! (but if you add 0.1 TWO times the answer is 0.2, which is correct - meaning python is not consistent when it is wrong either and you'll never know when python gives you the correct answer - yayy!). The no paranthesis but indentation stuff is a nice idea but EXTREMELY difficult for children to understand. If the child makes ONE indentation out of place python spits out strange messages NOT related to the indentation error, but side effetcs of the now flawed logic.
@@curiousLeafy naw. 250 years is hella long. How many programming languages and risen and fallen in the last 50 years? What makes you think that python will last 200+ years?
8:35 strange seeing you being so positive on 'alien perspective'....from the perspective of aliens located at, say, 1000 ly away,they'll see the following timeline: 10000 bc(9000 bc in our reference frame) : apes start farming and changing landscape noticably -1500 ad(while we live in 2500 in our reference frame) : apes are prosperous and use dynamite(which makes flashes of anomalous spectrum) 1900 ad(2900): apes now use radio and start building flying machines and skyrocketing various explosives 1950(2950) : oh shit they've got nukes and spacecraft fokin fast , anomalous spectrum super high power density bombing, polluting their own habitat with radioactive isotopes. 2000s ad: (we are having year 3000 by that time) crazy ass apes devastating their own ecosystem (which is clearly seen via spectrum), lots of spacecraft, oxygen plummeting, wars all over the globe, nonsense transmissions like politics tv programmes and ads). no signs of being able to 1) care much about their own survival long term 2) compassion towards other species 3) look at themselves from another perspective 4) avoid bad/rogue AI outbreak considering a) computation power progress b) aircrafts still crash regularly due to software quality. Aliens then ask themselves: "what the hell is happening with apes RIGHT NOW (they can see us with 1000yrs LAG but they know it) , how far has apes' technology gone, how dangerous have they become from 2000 to 3000 in APES' REFERENCE FRAME while we watch them from 1k ly away and HADN'T THEY ALREADY STARTED their low quality HIGHLY ARMED BADLY CONTROLLED drone swarms AIDED aggressive expansion to, say, produce more bacon and luxurious shit and bitcoins? should we fire our Halo-like sanitizing full spectrum entropizing guns towards apes just in case apes have already been approaching our area for 500+ years while we only received their 2000ad events via slow photons? (remember-aliens know it's 3000ad for us while they watch our 2000s). considering the rate of technical progress of apes....
so imagine our civilization in 2300, looking back at 1900-2000. we will see a stupid aggressive attention-deficient group of passengers on our space raft. so....then it turns out that we have data about only 1 biota and it's evolutionary and technological progress (ourselves). so we must(I mean we MUST due to mathematical statistocal laws of inference) consider any other newborn civilization to have about the same flaws/aggressiveness/shortsighted policy and resources-usage-wide stupidity and AI-self-assembled drones production capabilities. what should we do in order to maximize survival chances in case of same stupid+aggressive+techno-advancing biota emerging near us? on such a strategic meeting, I'll vote for FOKIN HIDE ASAP FIRST option. and you? remember: right now a sphere of photons containing visual of our vessel personnel actions is expanding in all directions through Laniakea, where a number of possible covert bad-biota sensors/radars covered in metamaterials are located, just fokin waiting to capture emissions of anomalous spectrum from nukes made by another dirty short sighted biota.
and yeas, in the long run we're gonna need such sensors mesh installed too for security purposes. FOR EXAMPLE who knows what will happen when one of Voyagers, full of bacterial/fungal/viral shredded rna/dna (some of which can likely (considering the consequences risk magnitude) turn out to be AUTOCATALYTIC, just like the one naughty molecule 4kkk years ago. this shit happens occasionally you know. ) hit some warm salty ocean some 1-100k ly away, giving birth to another biota unintentionally. YES chances are low everyday-life-wise and no insurance agent is now willing to take this risk into account. but just look at the trend of how risks are evaluated and perceived history-wise, and realize that we're talking existential risks here.
Man after I started learning OOP (only in Python, sorry ahah), I feels like us brain and also stomach and other organs are a type of object, and each fucking part of us body have a class, that sends information to the prefrontal cortex (that is the script), so execute everything
@@alimations9284 Natural language is full of ambiguity. What does "“The man saw the girl with the telescope” mean? Did the man use a telescope to see the girl, or did he see a girl that had a telescope?
I don't think so since how many developers Rust has got vs over 8 million Python developers? Whatever is popular would continue longer than Rust and others... The number of Python developers is still growing... but what about Rust?
Julia is the one you should watch out for. MIT created Julia specifically to address the limitations of Python without sacrificing on readability. Rust will however slowly replace C/C++ in low-level programming.
In the far future there will be only two programming languages. Javascript and Python. Javascript for frontend and Python for backend. Everything will be a web app.
Then people would still try to create better web servers in fast memory safe languages like Rust. Then maybe better alternative to javascript or maybe something completely different from a browser. Can't really say about the future.
Who is? Gudio? You know he built Python right? or Lex? You know he is a Ph.D. computer scientist? I think both of them are pretty far from "clueless" lol.
Nikola tesla, George Westinghouse, Edison could be listened to for days on end so there is no excuse. Guido needs to work on doing interviews or he can program an AI bot to do his interviews for him.
Full podcast episode: th-cam.com/video/-DVyjdw4t9I/w-d-xo.html
Lex Fridman podcast channel: th-cam.com/users/lexfridman
Guest bio: Guido van Rossum is the creator of Python programming language.
Just graduated with a degree in software engineering. We still had to learn how to make a basic processor and made a pong game from just binary inputs. The fundamentals are still being taught. They’re just being forgotten quickly with trying to learn all the new frameworks and dev tools out there.
Which, from an educational standpoint, is fair enough. People do tend to forget what they either don't use or don't find interesting, though the knowledge might still be usefull for someone.
hey everybody, this guy JUST GRADUATED!... See? noone cares!!!... jk congrats on the degree big boss
Man, this is very cool, but is it normal to do something like this in college?
I think there’s a middle ground. I appreciated the low level concepts I learned in UC schooling. It’s helped.
I’m also pretty sure I didn’t have to program battleship in assembly (for two players where it’s human v human or human v computer) to really understand low level programming.
Could have done without that suffering.
Congrats on the degree. The problem with prioritizing learning these lower-level parts is it holds back and gatekeeps technological advancement. We need people that can do the job, not necessarily build the tools to do the job. Those are specializations and separate jobs now. People train in these specializations but the more foundational people wont hire them and stuff stagnates.
Python will be around for a while. I don't know who is even maintaining gcc at this point. But it's still being maintained, because it's critical.
You need C to make python. Python is a C/C++ pre-compiled moduled based system. No C, no python.
@@colinmaharaj50 well C will live long past python so that won’t be a problem. I struggle to imagine a future of computing that isn’t extremely dependent on C.
You can look up the GNU Project's page on GCC. It has a steering committee that makes decisions for the project. And the organization behind GCC -- the GNU Project -- is the same as it's always been.
@@colinmaharaj50 Various other implementations of Python exist. My calculator has some non-CPython running on it, and I know off the top of my head that Jython and IronPython (C#) exist
@@samdcbu well we should consider it C is quite hard to develop in at times it could be nice if we ever do to C as we sort of did with assembly and make a language that compiles extremely efficiently to C but with higher capabilities and no python doesn’t really count for that
Let us pray JavaScript becomes legacy long before Python.
I think webassembly is probably be the js killer in the future
I love python. It feels natural to code in it. Almost like writing a text in natural language
The thing that actually feels like that - in an unpleasant way, I might add - is Kotlin. It feels like speaking English except with an unusual form of Tourette's syndrome where you say some other punctuation like periods and brackets in the middle. It's actually disorienting! Python on the other hand is friendly without being patronizing
you probably never heard before about Ruby right?
@@achrafaguenaou7825it’s kind of obvious that he was saying that it’s easier to get a job as a front end dev compared to the difficulty of getting any job with python. Backend, data science, whatever it may be- he’s saying it’s harder to get a job in those fields without a degree than it is to get a job as a front end dev. I felt like what he said was relatively straightforward so I don’t know how else to boil this one down.
@@jamespower5165 IMHO Kotlin does functional programming better than Python.
1:40 Yeah, same here, I started with transistors and hand made digital circuits in 1991. My first digital circuit was a ramp generator to make a Digital multimeter. I did 8085 ASM, then I got a 286 when I was employed in my 1st (or 2nd) year. In office they had a 8086 IBM XT so I was more highly advanced at home. LOL.
At some point it will be hard to find a compromise between adding more features to the language, maintaining backwards compatibility and preserving simplicity, and being simple is one of the core design points of Python. I believe that is one of the reasons Python 2 and 3 are not compatible, they decided it was a more acceptable trade-off. But it can't continue forever, it will either stop or Python become new C++. Of course there will be new languages inspired by Python.
We are going to need much bigger memory first on our micro-controllers for Python to be the new C++
Julia is one of those languages, it was created to address Python's speed limitations without compromising on readability.
@@wyqtor mojo also exists now
@@aoeu256 Yes, I am following Mojo as well. Very promising.
This lecture second half is the Gem which I want to uncover in my lifetime.
I actually don't think he's right here, and I think people who think this endless abstraction can continue are the same people who think that the world financial system can't collapse. We keep just stacking more shit, more layers of abstraction, but if we do that for a long enough time, then we're eventually going to have systems where not only does no one know actually how they work (which we have now), but we'll have entire systems where it will take multiple life times of research just to figure out how they work.
Machine learning isn't the answer, because that's just adding more layers of abstraction--and obfuscated abstraction at that--which will hide how things work. Even worse, if it really can program beyond human capability, then we're not going to be able to understand the code it output, and as a result how can we audit it? We'll have to just "trust" the ai, which I think is not a good idea.
I really feel like we're building the tower of babel, and we're headed for a very very big fall here.
Very insightful comment.
This is probably why colleges start from the basics of computer science wrt computer architecture, digital logic, operating systems... Probably down the line their curriculum could also become abstracted.. Which is a terrifying thought to me.
but maybe future humans can live far longer than a century
The more complex the system the more complex the abstractions. Complexity causes more abstraction. If complexity of what software should do increases while no more abstraction layers are added it will be harder, not easier to understand how everything works compared to the situation where the rise in abstraction follows the rise in complexity.
Your comment reminded me of a guy who claimed the busses cause rush hour. During the rush hour when people are getting back from work there are more buses on the street. When rush hour is over the number of buses on the street decreases. He concluded that buses are the cause of traffic. In reality, buses help make the traffic less bad and if the rise in traffic isn't followed by the rise in buses on the street, the traffic would be much worse.
Oh, I have just noticed that OP is the idiot who claimed C is not pass by value on a different video. Yeah, this is the kind of "expert" that we are dealing with. He actually isn't any better than that guy who I met on the bus who argued that the traffic is so bad because there are too many buses on the streets. Read my previous comment for context
I’m a data analyst at a big streaming company. We need faster languages than Python, SQL and R as datasets get ever larger. I work with pretty large datasets and building intermediary tables, optimising and sampling can only take you so far. These languages are getting inadequate for modern computing needs - especially with regards to data. Interested to see what new language supersedes them…
Did you alrdy try Julia?
Even native Python will be 5 times as fast in 5 years. You'll probably have better C/C++ bindings, faster algorithms written into libraries, better multithreading support - and faster computers as well on the hardware side of things. Ultimately library support matters and that's where Python will stay unbeatable
@@a.7388 It sounds nice but it's a buggy mess.
Rust?
What about Go? Would it have use cases for you?
On replication, fascinating thoughts. The interesting thing is that we are not generally aware of the layers upon layers of things that must fall in place for nature to create. And it's a miracle that it works so flawlessly almost everytime.
van Rossum is a wonderful human-being. Thanks for doing this series of interviews, I've really enjoyed them.
python took out perl, something will take out Python
Python took out BASIC.
Guido Von Rossum: "We already got one!"
Obviously eventually it will be a legacy language, but the thing is, what will replace it and be easier to use?
Python is one of the easiest and most supported languages to use right now, I don't see any other language replacing it for what it does unless you need some extremely efficient and fast code or a front-end web app.
50 - 100 years is an extremely long time in software. Consider how prevalent mobile applications have become in just 15 years. I agree that Python is currently in a great spot, but long-term, there's no telling what the future of software holds.
The good news is that the only way for it to be a legacy language is if a new and easier language comes around. And when that happens, it'd be easy for current python users to move along. Just like how easy it is for C developers to pick up python.
@@tayern3927 yeah, there is no problem at all. Better tools easier job.
@@raenastra
The further polished an idea gets the higher the longevity, if not for planned obsolescence.
Julia: *Heavy Breathing*
Thanks
He makes a point clearly..might be a legacy one but can not be replaced.also it will take at least 50 + years for the conversion..and also will be so only when bots chooses the tech stack
But what a legacy it will be... :)
Musicians with modular synths play with basic (CMOS 4xxx) logic daily... 🤷♂️ It is still fully useable, specially in rhythmical pattens 8 or 16 note patterns are very common all the time.
lol, software made on top of something unkown (like a programmer that does not how a cpu works), and those who make the cpu and know how it works, don't know in the end what electricity is in the end. Amazing
I personally believe that C/C++ will always remain the elite of all programming languages.
Rust will probably overtake C++ eventually, but even that will take at least another generation or so.
@@mattreigada3745 at least 100 years.
@@gonkong5638 won't be that long. Biggest deciding variable is development and widespread adoption of open source alternatives to the Linux kernel. That is unlikely to ever be needed, so C is safe, but C++ is more of an open question.
Either Rust overtakes C++ within maybe a generation, or something shinier comes along cutting Rust down before it has a critical mass of projects. Unlike C, C++ has no real central nucleus of development though, so popularity is more variable with support for frameworks and language bindings.
@@mattreigada3745 Yeah you're right, I was just exaggerate. But I think C would survive pretty long.
Yes. Especially for video game development and any other piece of software where performance is critical.
yet another layer of abstraction....
Crash course computer Science by PBS studios, Got ya.
Damn so there’s gonna be a new one in the future that will lead the job market?
Definition of a legacy language:
A ProgrammingLanguage which is still used in production code, but only in LegacySystems. In other words, where the language is typcally only selected when its use is a bona fide requirement, usually due to the need to interface with zillions of lines of code that already exist.
Eventually people will become legacy too
Who designed that schematic? Where is the creator?
So what should I do, I am in cybersecurity and I am now just learning about user input in python, should I stop learning the course?
Learn rust or golang
I think python is worth the learn. Let say Python is to slow for your use case so you make the real computation using Cpp or Rust, you can use python to call those package
I'm not invested in cybersecurity but I am a software engineer - Python will be around for a while, long enough at least for you to get familiar with coding and syntax in general. Once you're more well versed in writing programs for your use case then it will be much easier to switch to new languages need if needed and tackle more tough solutions.
Honestly it doesn't matter much the language that you start with, but a simple one like python helps to focus on the important things
No, you should definitely continue learning Python.
Am I imagining it or did Guido look pissed at the GIL joke?
I like lex taking notes dats how I learn syntax
epic interview ultimate guest up there with musk and chamath big thanks lex you are doing a g r e a t job and service for a l o t of folks!
hey, don't compare Guido, an actual improvment to society, to m*sk
Python for me is just to slow. That said I understand that it's taught in schools and is populair now. But vb5 was super populair when I was a kid/ teenager. These things often change over time quite quickly
Python is only slow if you're using it for the wrong purposes. The speed critical stuff should be compiled C/C++ code that can be imported into Python... which could then act as a general interface to it. This is the reason why Python is doing well in the SciComp community.
Right now the programming philosophy is easy to write, easy to read code, and python is extremely good for this.
I don't know what would replace it.
A new language that easy to read and write like python, fast like c, memory safe like rust, and dont forget about versatility and cross platform.
@@nasimicin Look at "Go" ;-)
It's a little bit simplistic so some people who want the complexity of C++/Rust are disappointed.
But it's quite simple and it compiles to machine code like C/C++/Pascal/COBOL/Fortran
Another example would be "Julia" - a "faster Python", basically.
Python is a very good and simple language
have you tried running a for loop on python
@@ImperialArmour depends on the fisherman capacity to learn it...
@@ImperialArmour Are you talking about speed?
@@jamespower5165 yup
It is simple for basic stuff, then when you start using properties and ABC it just gets awkward. It was simple, then it got the walrus operator.
Lex Fridman == Ian Malcolm
kinda sounded like he meant programming in general will be legacy, IMHO py might 1 of the last to stand
Did they smoke weed before this interview?
Nim will take Python
not sure , nim is good but we need a scripting language for some web/data science tools , i don't know why we still making more general use programming languages ,
C/C++ , go , python , js and rust could be the only languages that we use and other people just commit to them instead of creating new ones i can even delete go from the list , but they build a huge ecosystem around it (docker , kubernetes...)
don't get me wrong i like trying languages , but a lof of time i don't really see the need
tried D,V,Nim, julia ,peregrine , but i did not see a lot of added value
Larry Wall... when?
nice interview but it feels like two mates doing pot :D
Isn't this what Java is now?
Why trolls attack Guido lmao
Of course, I am not a futurologist but I believe it is hard to find something more inspiring than C++ running on Linux. All other languages are sort of derivatives or shortcuts of the most probably powerful language human being pioneered. Sorry, this is only my humble point of view.
But C++ is just "C with fancy classes". And C is just "B but with less limitations".
There was D as a "modern C++" but it failed by being just a little bit better so there is no real reason to use it.
Maybe Rust would be different enough to become that "next C++"? ;-)
@@igorthelight You are right. I think however that Carbon will extend C++ abilities since it is back compatibility with C++. Anyway, I do not see the reason why people have to resign from using C++. The beauty is the fact that in CS everyone can find something which is the most suitable for solving a particular problem. I do not have anything against any language and I would like to avoid saying one language is better than the other. Good luck.
Assembly is faster but yes C++ strikes the best balance. Different languages serve different purposes. C, C++ are still better when speed of execution is desirable. Python is better from the view that you don't even need to be a professional programmer to write code in it and so a vast array of libraries gets built up
@@igorthelight "But C++ is just "C with fancy classes"."
That hasn't been true for 40 years.
@@isodoubIet I know that it keeps evolving and the gap is widening ;-)
I think it's pretty naive to compare the human brain to a computer. Throughout history, humans have always compared the function of the brain to whatever the most advanced technology of the time was. People used to say the brain was like a clock or a watch. Now they think of it in terms of computers and code.
The fact is, we know very little about how the chemistry of the brain translates to actual thought and perception, and comparisons to computers are just speculation based on existing technology we understand. Brain function, while we know some things, is still largely undiscovered territory. We don't know if machine learning / deep learning are really similar at all to brain function at all because we simply do not understand brain function. Machine learning can give the appearance of thought in some cases, but it doesn't really act like a human at all, and for the most part, it is very good a some things humans are bad at and very bad at some things humans are very good at.
It's not even clear to me that humankind has the capacity to comprehend the mechanism of our own thought. Maybe we do, but we aren't anywhere close to that point yet.
The way they are pushing Python to non technical jobs, no wonder this is more than a tech.
Sadly python is flawed and not easy at all to understand for children. Try adding 0.1 three times like this: x = 0.1 + 0.1 + 0.1 THEN print( x ) gives 0.30000000000000004 , yayy!
(but if you add 0.1 TWO times the answer is 0.2, which is correct - meaning python is not consistent when it is wrong either and you'll never know when python gives you the correct answer - yayy!). The no paranthesis but indentation stuff is a nice idea but EXTREMELY difficult for children to understand. If the child makes ONE indentation out of place python spits out strange messages NOT related to the indentation error, but side effetcs of the now flawed logic.
Use Python Decimal.
The future they are talking about is nearly 200-300 years in the future but before that learn python if you want to.
200-300 years ? Lul
Humanity went from fighting with metal sticks to launching nukes in 200-300 years.
@@goochipoochie let's finalize in middle atleast 250 years
@@curiousLeafy naw. 250 years is hella long. How many programming languages and risen and fallen in the last 50 years? What makes you think that python will last 200+ years?
8:35 strange seeing you being so positive on 'alien perspective'....from the perspective of aliens located at, say, 1000 ly away,they'll see the following timeline:
10000 bc(9000 bc in our reference frame) : apes start farming and changing landscape noticably
-1500 ad(while we live in 2500 in our reference frame) : apes are prosperous and use dynamite(which makes flashes of anomalous spectrum)
1900 ad(2900): apes now use radio and start building flying machines and skyrocketing various explosives
1950(2950) : oh shit they've got nukes and spacecraft fokin fast , anomalous spectrum super high power density bombing, polluting their own habitat with radioactive isotopes.
2000s ad: (we are having year 3000 by that time) crazy ass apes devastating their own ecosystem (which is clearly seen via spectrum), lots of spacecraft, oxygen plummeting, wars all over the globe, nonsense transmissions like politics tv programmes and ads). no signs of being able to 1) care much about their own survival long term 2) compassion towards other species 3) look at themselves from another perspective 4) avoid bad/rogue AI outbreak considering a) computation power progress b) aircrafts still crash regularly due to software quality.
Aliens then ask themselves: "what the hell is happening with apes RIGHT NOW (they can see us with 1000yrs LAG but they know it) , how far has apes' technology gone, how dangerous have they become from 2000 to 3000 in APES' REFERENCE FRAME while we watch them from 1k ly away and HADN'T THEY ALREADY STARTED their low quality HIGHLY ARMED BADLY CONTROLLED drone swarms AIDED aggressive expansion to, say, produce more bacon and luxurious shit and bitcoins? should we fire our Halo-like sanitizing full spectrum entropizing guns towards apes just in case apes have already been approaching our area for 500+ years while we only received their 2000ad events via slow photons? (remember-aliens know it's 3000ad for us while they watch our 2000s). considering the rate of technical progress of apes....
so imagine our civilization in 2300, looking back at 1900-2000. we will see a stupid aggressive attention-deficient group of passengers on our space raft. so....then it turns out that we have data about only 1 biota and it's evolutionary and technological progress (ourselves). so we must(I mean we MUST due to mathematical statistocal laws of inference) consider any other newborn civilization to have about the same flaws/aggressiveness/shortsighted policy and resources-usage-wide stupidity and AI-self-assembled drones production capabilities. what should we do in order to maximize survival chances in case of same stupid+aggressive+techno-advancing biota emerging near us? on such a strategic meeting, I'll vote for FOKIN HIDE ASAP FIRST option. and you?
remember: right now a sphere of photons containing visual of our vessel personnel actions is expanding in all directions through Laniakea, where a number of possible covert bad-biota sensors/radars covered in metamaterials are located, just fokin waiting to capture emissions of anomalous spectrum from nukes made by another dirty short sighted biota.
and yeas, in the long run we're gonna need such sensors mesh installed too for security purposes. FOR EXAMPLE who knows what will happen when one of Voyagers, full of bacterial/fungal/viral shredded rna/dna (some of which can likely (considering the consequences risk magnitude) turn out to be AUTOCATALYTIC, just like the one naughty molecule 4kkk years ago. this shit happens occasionally you know. ) hit some warm salty ocean some 1-100k ly away, giving birth to another biota unintentionally. YES chances are low everyday-life-wise and no insurance agent is now willing to take this risk into account. but just look at the trend of how risks are evaluated and perceived history-wise, and realize that we're talking existential risks here.
Next programming language won't even be one. Age of promts is about to come.
I think that such a scenario would take decades to happen...
guido jonguuuuh, lekker bezig pik.
We need to get rid of fracking JavaScript and HTML
What? HTML?
@@eeriemyxi That's right. the current web "technology" is full of crap. It just happened. Never really designed.
they sound like they are getting higher and higher as this interview progresses
mostly Lex though :D
Did Lex really bring up MetaVerse? Really? Can he explain how it would actually help the future generations? Stop hanging with Zuck.
Man after I started learning OOP (only in Python, sorry ahah), I feels like us brain and also stomach and other organs are a type of object, and each fucking part of us body have a class, that sends information to the prefrontal cortex (that is the script), so execute everything
The only thing simpler than python is using natural language to specify programs, which is (fortunately) still not a thing
Fortunately? why? Are you afraid of losing your job as a programmer?
@@alimations9284 Natural language is full of ambiguity. What does "“The man saw the girl with the telescope” mean? Did the man use a telescope to see the girl, or did he see a girl that had a telescope?
I wonder if a language like Rust could replace Python and Java in universities. It's certainly an interesting language!
I don't think so since how many developers Rust has got vs over 8 million Python developers?
Whatever is popular would continue longer than Rust and others...
The number of Python developers is still growing...
but what about Rust?
Julia is the one you should watch out for. MIT created Julia specifically to address the limitations of Python without sacrificing on readability. Rust will however slowly replace C/C++ in low-level programming.
@@wyqtor
Rust has a long long way to go...
Maybe after 2 decades to replace anything...
It looks like Lex is having a hard time pronouncing all that complicated words
In the far future there will be only two programming languages. Javascript and Python. Javascript for frontend and Python for backend. Everything will be a web app.
One can only hope.
If Javascript has Typescript, what will "enhance" python?
Then people would still try to create better web servers in fast memory safe languages like Rust. Then maybe better alternative to javascript or maybe something completely different from a browser. Can't really say about the future.
Web assembly will eventually kill javascript. Javascript is a terrible language.
@@PythonPlusPlus Python runs in the browser with PyScript. It's only a matter of time. Probably not much time.
He never really answer the questions, always go on a his own new topic..*sigh*
he is clueless
Lmao
probably more clueful than you can ever hope be 😅
Who is? Gudio? You know he built Python right? or Lex? You know he is a Ph.D. computer scientist? I think both of them are pretty far from "clueless" lol.
no wonder python started off as a mess
it was written by a fool
Why would you say this?
@@avithedev Lmfao, what is happening in this comment section.
@@avithedev you mean you watched the whole video and couldn't tell why?
I'm the worst programmer, but it seems easy to hack.
@Gyrate he is the worst programmer for a reason lol
Garbage collection. Yes. That's why Rust has its different philosophy for cybersecurity. Python is mainly used in datascience and web development
This guest went straight to legacy without current appreciation because he is hard to watch and listen to.
the disrespect
What the he’ll am I reading?
Disrespect deserves none appreciation though...
Nikola tesla, George Westinghouse, Edison could be listened to for days on end so there is no excuse. Guido needs to work on doing interviews or he can program an AI bot to do his interviews for him.
@@BrooklynNYC777 why? Why does he need to get better at interviews? Its not like thats his livelihood
Who is the gay guy?