I think Mojo's cool and there's definitely a place for a modern applications programming language with C++ class(ish) performance, aka what Swift wanted to be but got trapped in the Apple ecosystem (designed by the same person as Mojo).
The strong AI focus seems to be a sign of the times, and not actually something that makes sense imo.
You don’t raise $130M at a $600M valuation to make boring old dev infrastructure that is sorely needed but won’t generate any revenue because no one is willing to pay for general purpose programming languages in 2025.
You raise $130M to be the programming foundation of next Gen AI. VCs wrote some big friggen checks for that pitch.
To my naive mind, any language that is controlled by a single company instead of a non profit is a non-starter. Just look at how many companies reacted when Java license change happened. You must be either an idiot or way too smart for me to understand to base your business on a language like Mojo instead of Python.
Anyone knows what Mojo is doing that Julia cannot do? I appreciate that Julia is currently limited by its ecosystem (although it does interface nicely with Python), but I don't see how Mojo is any better then.
I've looked into making Python modules with Julia and it doesn't look like that is very well supported right now. Where as it's a core feature of Mojo.
Weird that there has been no significant adoption of Mojo. It has been quite some time since it got released and everyone is still using PyTorch. Maybe the license issue is a much bigger deal than people realize.
I'm on the systems side, and I find some of what Chris and team are doing with Mojo pretty interesting and could be useful to eradicate a bunch of polyglot ffi mess across the board. I can't invest in it or even start discussions around using it until it's actually open.
They’re not going to see serious adoption before they open source. It’s just a rule of programming languages at this point if you don’t have the clout to force it, and Modular does not. People have been burned too many times by closed source languages.
I definitely think the license is a major holdback for the language. Very few individuals or organisation for that matter would like to invest in a new closed stack. CUDA is accepted simply because it has been along for such a long time. GPGPU needs a Linux moment.
The market tends to be pretty efficient for things like these. We’ve seen significant rapid adoption of several different ML solutions over the last decade, yet Mojo languishes. I think that’s a clear sign they aren’t solving the real-world pain points that users are hitting, and are building a rather niche solution that only appeals to a small number of people, no matter how good their execution may be.
Mojo is the enshitification of programming. Learning a language is too much cognitive investment for VC rugpulls. You make the entire compiler and runtime GPL or you pound sand, that has been the bar for decades. If the new cohort of programmers can’t hold the line, we’ll all suffer.
What are you ranting about? Lattner has a strong track record of producing valuable, open source software artifacts (LLVM, Swift, MLIR) used across the industry.
When I was young, I enjoyed messing around with new languages, but as time went on, I realized that there is really very little to be gained through new languages that can not be obtained through a new library, without the massive downside of throwing away most of the ecosystem due to incompatibility. Also, CuPy, Triton and Numba already exist right now and are somewhat mature, at least compared to Mojo.
Listening to this episode, I was quite surprised to hear that even now in Sept 2025, support for classes at all is considered a medium-term goal. The "superset of Python" angle was thrown around a lot in earlier discussions of Mojo 1-2 years ago, but at this rate of progress seems a bit of a pie-in-the-sky aspiration?
The reason why Python dominates is that modern ML applications don't exist in a vacuum. They aren't the standalone C/FORTRAN/MATLAB scripts of yore that load in some simple, homogeneous data, crunch some numbers, and spit out a single result. Rather, they are complex applications with functionality extending far beyond the number crunching, which requires a robust preexisting software ecosystem.
For example, a modern ML application might need an ETL pipeline to load and harmonize data of various types (text, images, video, etc., all in different formats) from various sources (local filesystem, cloud storage, HTTP, etc.) The actual computation then must leverage many different high-level functionalities, e.g. signal/image processing, optimization, statistics, etc. All of this computation might be too big for one machine, and so the application must dispatch jobs to a compute cluster or cloud. Finally, the end results might require sophisticated visualization and organization, with a GUI and database.
There is no single language with a rich enough ecosystem that can provide literally all of the aforementioned functionality besides Python. Python's numerical computing libraries (NumPy/PyTorch/JAX etc.) all call out to C/C++/FORTRAN under the hood and are thus extremely high-performance, and for functionality they don't implement, Python's C/C++ FFIs (e.g. Python.h, NumPy C integration, PyTorch/Boost C++ integration) are not perfect, but are good enough that implementing the performance-critical portions of code in C/C++ is much easier compared to re-implementing entire ecosystems of packages in another language like Julia.
> There is no single language with a rich enough ecosystem that can provide literally all of the aforementioned functionality besides Python.
Have a hard time believing C++ and Java don't have rich enough ecosystems. Not saying they make for good glue languages, but everything was being written in those languages before Python became this popular.
I'm in kind of a different place with it on the inference side.
I've got these crazy tuned up CUDA kernels that are relatively straightforward to build in isolation and really where all the magic happens, and there's this new CUTLASS 3 stuff and modern C++ can call it all trivially.
And then there's this increasingly thin film of torch crap that's just this side of unbuildable and drags in this reference counting and broken setup.py and it's a bunch of up and down projections to the real hacker shit.
I'm thinking I'm about one squeeze of the toothpaste tube from just squuezing that junk out and having a nice, clean, well-groomed C++ program that can link anything and link into anything.
Ironically Python is the worst language for everything you’ve described. Packaging is pain, wheels are pain, everything breaks all the time. It’s only great for those standalone scripts. Nobody in their right mind would design Python the way it turned out if the goal was to be the main ML language.
You argument is circular. Python has all this ecosystem _because_ it have been the language of choice for ML for a decade. At this point it's difficult to beat, but doesn't explain why it was chosen all those years ago.
I was there when Perl and Tcl were the main actors, that is why VTK used Tcl originally.
Python dominates, because 25 years ago places like CERN started to adopt Python as their main scripting language, and eventually got used for more tasks than empowered shell scripts.
It is like arguing why C dominates and nothing else can ever replace it.
Thank you for all the great interest in the podcast and in Mojo. If you're interested in learning more, Mojo has a FAQ that covers many topics (including "why not make Julia better" :-) here:
https://docs.modular.com/mojo/faq/
I don't think ML does need a new programming language. You give up an extreme amount of progress in tools and libraries when you move to a new language.
I haven't seen new languages that market themselves for specific features that couldn't be done just as easily through straight classes with operator overloading.
I'm the primary target audiance for Mojo and was very interested in it, but I just wish they didn't keep Exceptions. This backwards compatibility with Python syntax is extremely overrated and not worth the cost of bringing language warts from the 90s.
God, I hate exceptions so much. I have never seen anyone use exceptions correctly in either Java (at FAANG) or in any regular Python application.
I'm much more in favor of explicit error handling like in Go, or the syntax sugar Rust provides.
I know it's a bit of trope to say this on HN... but why not Lisp?
If I make the assumption that future ML code will be written by ML algorithms (or at least 'transpiled' from natural language), and Lisp S-Expressions are basically the AST, it would make sense that it would be the most natural language for a machine to write in. As a side benefit, it's already the standard to have a very feature complete REPL, so as a replacement for Python it seems it would fit well
The mojo faq talks about the language as if it's a strict superset (or aiming to be) of Python. Or that mojo "is" Python.
Yet the roadmap says:
> As Mojo matures through phase 3, we believe Mojo will become increasingly compatible with Python code and deeply familiar to Python users, except more efficient, powerful, coherent, and safe. Mojo may or may not evolve into a full superset of Python, and it's okay if it doesn't.
This is incredibly confusing. If it's _not_ aiming for Python compatibility, why are we talking about Python at all?
Also, is anyone actually considering using an emoji as a file extension?
51 comments
[ 2.0 ms ] story [ 65.7 ms ] threadThe strong AI focus seems to be a sign of the times, and not actually something that makes sense imo.
https://techcrunch.com/2023/08/24/modular-raises-100m-for-ai...
You don’t raise $130M at a $600M valuation to make boring old dev infrastructure that is sorely needed but won’t generate any revenue because no one is willing to pay for general purpose programming languages in 2025.
You raise $130M to be the programming foundation of next Gen AI. VCs wrote some big friggen checks for that pitch.
For example, a modern ML application might need an ETL pipeline to load and harmonize data of various types (text, images, video, etc., all in different formats) from various sources (local filesystem, cloud storage, HTTP, etc.) The actual computation then must leverage many different high-level functionalities, e.g. signal/image processing, optimization, statistics, etc. All of this computation might be too big for one machine, and so the application must dispatch jobs to a compute cluster or cloud. Finally, the end results might require sophisticated visualization and organization, with a GUI and database.
There is no single language with a rich enough ecosystem that can provide literally all of the aforementioned functionality besides Python. Python's numerical computing libraries (NumPy/PyTorch/JAX etc.) all call out to C/C++/FORTRAN under the hood and are thus extremely high-performance, and for functionality they don't implement, Python's C/C++ FFIs (e.g. Python.h, NumPy C integration, PyTorch/Boost C++ integration) are not perfect, but are good enough that implementing the performance-critical portions of code in C/C++ is much easier compared to re-implementing entire ecosystems of packages in another language like Julia.
Have a hard time believing C++ and Java don't have rich enough ecosystems. Not saying they make for good glue languages, but everything was being written in those languages before Python became this popular.
I've got these crazy tuned up CUDA kernels that are relatively straightforward to build in isolation and really where all the magic happens, and there's this new CUTLASS 3 stuff and modern C++ can call it all trivially.
And then there's this increasingly thin film of torch crap that's just this side of unbuildable and drags in this reference counting and broken setup.py and it's a bunch of up and down projections to the real hacker shit.
I'm thinking I'm about one squeeze of the toothpaste tube from just squuezing that junk out and having a nice, clean, well-groomed C++ program that can link anything and link into anything.
Python dominates, because 25 years ago places like CERN started to adopt Python as their main scripting language, and eventually got used for more tasks than empowered shell scripts.
It is like arguing why C dominates and nothing else can ever replace it.
Sorry, couldn't resist.
https://en.wikipedia.org/wiki/Standard_ML
Mojo also has a bunch of documentation https://docs.modular.com/mojo/ as well as hundreds of thousands of lines of open source code you can check out: https://github.com/modular/modular
The Mojo community is really great, please consider joining, either our discourse forum: https://forum.modular.com/ or discord https://discord.com/invite/modular chat.
-Chris Lattner
I haven't seen new languages that market themselves for specific features that couldn't be done just as easily through straight classes with operator overloading.
- https://www.youtube.com/watch?v=pdJQ8iVTwj8
- https://open.spotify.com/episode/6flH0XxwdIbayoXTHOgAfI
- https://podcasts.apple.com/us/podcast/381-chris-lattner-futu...
He has two other episodes on the show:
- https://www.youtube.com/watch?v=nWTvXbQHwWs (Episode 131, October 18th, 2020)
- https://www.youtube.com/watch?v=yCd3CzGSte8 (Episode 21, May 13th, 2019)
God, I hate exceptions so much. I have never seen anyone use exceptions correctly in either Java (at FAANG) or in any regular Python application.
I'm much more in favor of explicit error handling like in Go, or the syntax sugar Rust provides.
If I make the assumption that future ML code will be written by ML algorithms (or at least 'transpiled' from natural language), and Lisp S-Expressions are basically the AST, it would make sense that it would be the most natural language for a machine to write in. As a side benefit, it's already the standard to have a very feature complete REPL, so as a replacement for Python it seems it would fit well
Optimal ML lang is half-way from LISP to APL.
Yet the roadmap says:
> As Mojo matures through phase 3, we believe Mojo will become increasingly compatible with Python code and deeply familiar to Python users, except more efficient, powerful, coherent, and safe. Mojo may or may not evolve into a full superset of Python, and it's okay if it doesn't.
This is incredibly confusing. If it's _not_ aiming for Python compatibility, why are we talking about Python at all?
Also, is anyone actually considering using an emoji as a file extension?