Full podcast episode: ua-cam.com/video/pdJQ8iVTwj8/v-deo.html Lex Fridman podcast channel: ua-cam.com/users/lexfridman Guest bio: Chris Lattner is a legendary software and hardware engineer, leading projects at Apple, Tesla, Google, SiFive, and Modular AI, including the development of Swift, LLVM, Clang, MLIR, CIRCT, TPUs, and Mojo.
Every time Lattner has worked on language infrastructure, he has never missed. Every language ecosystem *needs* an evolution strategy to get better. I'm optimistic that this will be Python's "successor".
From Pythonistas to Mojicians. I like it. Plus this has been done multiple times. Its good to move the industry forward by using better improves languages. Like: Objective-C -> Swift Java -> Kotlin JavaScript -> TypeScript C -> C++ -> (Carbon, Rust) And now Python -> Mojo.
idk about the c->c++ part, I also don't know about the c++ -> carbon part. If anything it seems like the market of C and C++ successors is going to be dispersed among multiple languages, and even then C is most likely going to perservere for a long time, on the other hand every person that has spent considerable ammount of time with c++ prays to God that some language finally succeeds that piece of bloated garbage
Wolfram... I encountered it in collge, but it was too expensive to use it. So I moved to Python and am very satisfied... But sometimes I miss the consistent rules and systems in Mathematica as it was maintained by one entity.
it could be more logical and marketing friendly because is the same sht they did with C, so if you know how have been things in the last 20 years, python++ were sure a better name.
Idk, I tried mojo. It significantly slowed down my python code, by about 20%. And was 600% slower than pypy. Maybe it’s good for certain things, but I think their marketing team overhyped it like crazy.
@@RJ-or8bw It was running 100% python code slower, didn't change any static typing. It was moreso that the language claims its a superset of python but it slows down actual python code, and is way slower than pypy code. Maybe there's specific use cases for it where that isn't the case, it just isn't marketed as such imo.
@@tseclassestseclasses4955 The industry is learn new things for the sake of learning new things though. I heard all these things already when he created Swift. Now, apparently, we need yet another new language. The truth is that these decisions are driven by personality not engineering.
Julia has shit libraries and it doesn't have the exact same syntax as python as a subset. The documentation is awful and there were/(and probably) are bugs in for instance zygote
Python is popular and known by a larger pool of people. Mojo also possesses features that are utilised in abundance in C++ and Java for example both of which have high exposure and usage in the world. It’s not complicated take the best of both worlds and put them into one as opposed to learning a brand new language
@@arthurpenndragon6434 he said not on 2023 or even 2024 he didn't said it is not coming but next 2025 the new update is coming and when that happens every system made with py 3 is gonna be outdate like the py 2
You definitely don't speak for the entirely of the Python user base. Being able to easily get features like SIMD/vectorization and a borrow checker while keeping the simplicity of Python's syntax, _and_ being compatible with all existing Python code, that has a lot of appeal to many.
Full podcast episode: ua-cam.com/video/pdJQ8iVTwj8/v-deo.html
Lex Fridman podcast channel: ua-cam.com/users/lexfridman
Guest bio: Chris Lattner is a legendary software and hardware engineer, leading projects at Apple, Tesla, Google, SiFive, and Modular AI, including the development of Swift, LLVM, Clang, MLIR, CIRCT, TPUs, and Mojo.
It's about being compatible with pre existing python code basis
This language should obviously have been named Monty and it is an unforgivable shame that is isn’t.
You are absolutely 💯 right
😂😂😂 right
Unforgivable ‼️
...And programmers in it, ([ 3:20 ]) "Monticians".
_Right?_
👑 You dropped this King
Every time Lattner has worked on language infrastructure, he has never missed. Every language ecosystem *needs* an evolution strategy to get better. I'm optimistic that this will be Python's "successor".
Great I heard some companies are hiring with 10 + years of experience for mojo
nah that's cap, I say I guy on LinkedIn who completed his Masters in 2022 or 23 that works at Modular
HAHA THIS JOKE NEVER GETS OLD
😂
Mojo for the win, As soon as it is out, I will go full out on it
I always love chris's language design
The field of programming languages is constantly evolving, I think.
It’s at least always uhhh changing
isn’t everything constantly evolving in its own way?
When will be Mojo fully out? Still in 2023?
This may surprise you, i still create ai in java, java 21 is so good, and future releases will be a game changer
Boomer..lol jk
Java is dead
Is this a meme account or real news? I'm interested if it's real news.
Any language with OOP paradigm only - has no future. Bc it is not OOP like Alen wanted
@@ITSecNEO "Good joke" :|
_- A 17 year old writing a game engine in Java in 2023._
2:21: To boldly go where no Python program has ever gone before.
I was imagining Python _in_ the kernel.
Mojicians, it is !!
Mojoheddin
It’s Mojitos. Obviously.
From Pythonistas to Mojicians. I like it. Plus this has been done multiple times. Its good to move the industry forward by using better improves languages. Like:
Objective-C -> Swift
Java -> Kotlin
JavaScript -> TypeScript
C -> C++ -> (Carbon, Rust)
And now Python -> Mojo.
Exactly! But industries are reluctant to adopting to new technology because they don't wanna recreate the entire framework which sucks
Its cool. But do not forget ABC and 1000+ another programming languages. Mojo is one of them.
Mojician 😂
@@BDOMyaccount ([ 3:20 ]) is where they call them that!
idk about the c->c++ part, I also don't know about the c++ -> carbon part.
If anything it seems like the market of C and C++ successors is going to be dispersed among multiple languages, and even then C is most likely going to perservere for a long time, on the other hand every person that has spent considerable ammount of time with c++ prays to God that some language finally succeeds that piece of bloated garbage
Well if things don't work out, Chris can always get a job as a diplomat!😂
Great communication skills!👍
So how does IDEs tool set and debugging work with Mojo?
It is still under active development.Around 8 Months old so the ecosystem is still developing
Amor
It's llvm based, ide tooling wil be quite easily to get up and running once it's being used in production
They provided a plugin for Visual Studio. JetBrains already demonstrated interest in a plugin for PyCharm or even a new IDE for it.
Still waiting for this
is there a plug in from Wolfram for Python ..
Wolfram... I encountered it in collge, but it was too expensive to use it. So I moved to Python and am very satisfied...
But sometimes I miss the consistent rules and systems in Mathematica as it was maintained by one entity.
Ther are WolframAlpha plugins for the Chat AIs so I'm assuming that under the hood these are being called from Python?
you can just call their api, if not you can call their web api
I also like python, my first language is javascript but i hate curly braces so i shift to python like language, the coffeescript.
C like programming languages forever 🏆
But Lex, where is your Flying V ?
They should have named it python++
it could be more logical and marketing friendly because is the same sht they did with C, so if you know how have been things in the last 20 years, python++ were sure a better name.
python+=1
Idk, I tried mojo. It significantly slowed down my python code, by about 20%. And was 600% slower than pypy. Maybe it’s good for certain things, but I think their marketing team overhyped it like crazy.
...Sorry!?
Dictionaries aren't optimized yet. Was it just running your 100% python code slower? Or did you add the static typing and it slowed you down?
@@RJ-or8bw It was running 100% python code slower, didn't change any static typing. It was moreso that the language claims its a superset of python but it slows down actual python code, and is way slower than pypy code. Maybe there's specific use cases for it where that isn't the case, it just isn't marketed as such imo.
Mojo isn't even open source. That's gross.
Not yet, they have plans to do so in the future once everything stabilises
Doesn't Julia solves this exact problem? I don't get it.
Well, jack of all trades, master of none.
@@tseclassestseclasses4955 The industry is learn new things for the sake of learning new things though. I heard all these things already when he created Swift. Now, apparently, we need yet another new language. The truth is that these decisions are driven by personality not engineering.
@alexshaykevich509 Agreed. High levels of egoism and sense of legacy with tech movers and shakers.
Julia has shit libraries and it doesn't have the exact same syntax as python as a subset. The documentation is awful and there were/(and probably) are bugs in for instance zygote
Python is popular and known by a larger pool of people. Mojo also possesses features that are utilised in abundance in C++ and Java for example both of which have high exposure and usage in the world. It’s not complicated take the best of both worlds and put them into one as opposed to learning a brand new language
"So...package ported to Mojo, then a Python interface." This is why Lex is the Feynman of computer science.
All my colleagues have moved away from python to golang. My friends in other companies are migrating to golang too. Too fast and easy to learn
Golang is not for AI.
Why golang and not rust?
@@armincal9834 easy and fast enough vs difficult...
this python craze must end. what an awful, awful thing to look at. hard to read hard to reason about. disguising. language design peaked at java
Smalltalk, Lisp（homoiconocity, macros, quote), also maybe the algebraic effect type languages seem much better designed than modern languages
Sorry but as fo me Mojo is a girl and boy in one body. Interesting? Yes. Wanna to be - not yet.
anime futanaries sound interesting, but you sure want girls and boys around anyway XD
I wish it didn't have python syntax, not a fun of those indentations
Python 4 is coming soon the game will change on u soon 🔜
No need to new Lang . We are good with py
guido himself said python 4 might not even happen.
There will be no python 4, there will be python 3.13, 3.14... 3.99, 3.100... 3.1488
@@arthurpenndragon6434 he said not on 2023 or even 2024 he didn't said it is not coming but next 2025 the new update is coming and when that happens every system made with py 3 is gonna be outdate like the py 2
After 3 years maybe because GIL is under removal with the help of meta engineer.
You definitely don't speak for the entirely of the Python user base. Being able to easily get features like SIMD/vectorization and a borrow checker while keeping the simplicity of Python's syntax, _and_ being compatible with all existing Python code, that has a lot of appeal to many.
He is lying