Anyway, never gonna try it, unless I'm absolutely forced. I hate their marketing strategy, it sounds too bad.
It's really minor thing but a pet peeve of mine, I remember how twenty years ago every new language claimed to be as fast or faster than C, and every time the benchmark was some ultra-optimized non-idiomatic program in that new language against a naive C implementation, and the C version was usually still 20% faster and use 10% as much RAM with the right compile flags.
Are there even Mojo benchmark in the wild? The FAQ entry[1] just finds excuses for situations where it does not perform well...
I would say "faster than C if you want to keep your sanity and don't have infinite time". It doesn't roll off the tongue though :P
- Run on GPU
- Automatic loop parallelisation
- Microbenchmarks where C is slow due to pointer aliasing (this is why Ocaml is faster than C)
[1] https://www.modular.com/blog/the-worlds-fastest-unified-matr...
I kind of agree with the obnoxiousness of the marketing, but it pays to actually read a little of what they do before jumping to conclusions.
If only because its Chris Lattner behind it, obviously not someone who is your average "modern programmer".
It's impossible to say for sure, since you can't run the benchmark on your machine and god knows what's behind the playground. However, I compared my naive implementation to find min of a rolling mean using python with numba, pandas, polars, numpy and mojo version.
Mojo version was faster than polars and numpy, which in turn were faster than pandas and numba. That's my N=1 of course, and not a reliable benchmark. But still, I find it very impressive.
When making technical claims to a technical audience, and vs Lattner leveraging reputation,
the appropriate approach is clearly not this. It's to lead with conservatively stated claims, reproducible results, and all the rest of the apparatus that is collectively understood as how adults (from which I exclude bizdev) demonstrate a better mousetrap, into an environment within which the cost of switching mousetrap technology is very great indeed at every level.
Big asks deserve and require better.
They are VC funded and this is the pitch to the VCs in order to get money.
Obviously this is not convincing on a technical level yet. But it's also really interesting as a design.
And while the code isn't open source, I for one find it pretty fascinating to see the language being iterated on in public like this.
Edit: E.g. https://github.com/modularml/mojo/blob/main/proposals/lifeti...
or
https://github.com/modularml/mojo/blob/main/proposals/value-...
In my experience the new generation of devs haven't touched C and treat it like some kind of weird arcane language, a bit like what Fortran and COBOL were when I started coding around 2000.
I doubt their potential users would care so much, as long as it's significantly faster than Python and they don't need C or C++, except as an option and not as a necessity. That's been the point of various other alternative languages.
The claim to fame of various newer compiled languages, is not just being as fast as C, but that they are better alternatives to use, for their audiences. Being easier to use, read, or/and learn. This is where some of the drama comes from, because some will make it about speed, and ignore the other elements of what is "attractive".
[1] https://en.wikipedia.org/wiki/History_of_programming_languag...
For performance, if the compiler backend performs enough optimization, the code is run natively without any VM, the performance can be same or better than C (as some language features may provide more optimization opportunities). At the end of the day, C programs are compiled into some IR code and optimized by the backend as well, so the frontend language is more about balancing limitation for optimization and expressiveness.
I do feel a bit uncomfortable about their marketing strategy, which compares a simple implementation against one with SIMD and parallelization... I think the real win here is simpler SIMD/parallel programming comparing with C/C++, instead of the language itself being faster.
"Faster than C" is a lot more plausible for Mojo than the claim of running arbitrary Python code eventually ;)
Indeed, I can see this from their emphasis on generic SIMD, which looks pretty nice. It is nice that there is now a language with first-class SIMD support, comparing with something implemented with a library or autovectorization which may not work.
> "Faster than C" is a lot more plausible for Mojo than the claim of running arbitrary Python code eventually ;)
Seems so. I don't recall reading the claim about running arbitrary Python code, and I think requiring rewriting Python code is reasonable as normal Python code may utilize runtime reflection/dynamic type and they are hard to implement efficiently, and probably require a GC for memory management.
The term superset of Python kind of implies this I think. Beyond the runtime reflection, Python also has eval.... :D
If the language is flexible enough (in particular, it has to support operator overloading), I don’t see any disadvantages to implementing that in a library.
See for example https://developer.apple.com/documentation/swift/simd
Except that it also claims to be a superset of Python. That is, you can use it to run any Python program without changes. So explain to me how it’s faster than C while also literally being Python. That’s more outrageous than V.
> what they've actually got at the moment is two languages glued together, where one of the languages will "unlock performance" and the other language is literally the CPython interpreter.
On the main page of the website, they claim it lets you "Write everything in one language" and "Unlock Python performance" – but what they've actually got at the moment is two languages glued together, where one of the languages will "unlock performance" and the other language is literally the CPython interpreter.
> Although not as outrageous as V...
Comes across as adding drama. Like arguing with one guy, then slapping another who wasn't involved, thinking it will be funny or because one might get away with it.
I know it was made for AI, but could it be a rival to rust/C++ in general? Seems like it could replace either of those.
Also, it is still WIP so I am not sure how things like type works in this language, whether or not it support parameterized lifetime, reference counted objects, union type, closure, etc. Currently, they seems to be focusing on numerical applications, targeting the AI crowd, so it is uncertain how things will go later.
This is very clearly explained on the website:
https://web.archive.org/web/20200629153446/https://vlang.io/
Can you point to that claim?
Here's 4 years ago:
https://web.archive.org/web/20190828170427/https://vlang.io/
> V manages memory at compilation time (like Rust). Right now only basic cases are handled. For others, manual memory management is required for now.
> V doesn't use garbage collection or reference counting. The compiler cleans everything up during compilation. If your V program compiles, it's guaranteed that it's going to be leak free.
This was later worded better. But I agree, the docs should have never had such wording.
People are also free to join the project or directly make helpful suggestions. That's part of the point of open-source.
(If I hear the randomly brought up, completely unrelated musk hype question again...)
He speaks slowly (listen on 2x for regular person speed) and he’s trite, but he interviews people and lets them talk.
That’s good enough. And it’s especially good enough in that he gets big names to come talk.
He doesn’t need to be a genius or to ask clever questions. Merely letting Chris Lattner geek out for 2 hours on audio about his project is a win.
His podcast is alright for people who aren't in any of the domains and therefore just want overviews. Anything else, and it's about as useful as listening to an actor repeat the summary for the movie they're promoting for the millionth time.
It's pretty clear he's playing up some kind of "rainman autistic AI savant" persona, and also, he exaggerates his credentials. On his website he describes himself as:
>AI researcher working on autonomous vehicles, human-robot interaction, and machine learning at MIT and beyond
Which could lead one to erroneously assume that he attended MIT, or even has a degree from MIT, when neither of which are true.
He had John Carmack on his show which piqued my interest, but he was asking questions like "what's the best programming language" and "what's the best IDE." Such a blown opportunity to be asking useless first year CS student questions like that when interviewing a great mind like Carmack.
It makes sense to get touchy-feely with Bert Kreischer, who’s basically an alcoholic gigantic third grader. Or with Michael Malice because they’re goofy best buds.
Still it’s nice when a serious guest kind of shrugs and sticks to the topic.
I can’t listen to Neil Degrasse Tyson talk because of that same kind of “the wonder of the poetry of the cosmic ballet” nonsense.
Fyi... CL said he likes being interviewed by Lex Fridman: https://news.ycombinator.com/item?id=24831157
So for your wish of someone else instead of Lex to be granted... (1) the hypothetical "more intelligent interviewer" you'd prefer has to ask CL to be interviewed -- and -- (2) in turn, CL has to say "yes" to the interviewer.
So far, Lex got (1) and (2) to happen. (Three times.) So here we are. Any suggestions on who else should interview CL and would CL grant the interview?
The More You Know...
[0] <https://lexfriedman.com/> [1] <https://en.wikipedia.org/wiki/Lex_Friedman> [2] <https://www.youtube.com/c/lexfriedman>
[3] <https://lexfridman.com/> [4] <https://en.wikipedia.org/wiki/Lex_Fridman> [5] <https://youtube.com/c/lexfridman>
This is a self-fulfilling prophesy.
"Please don't comment about the voting on comments. It never does any good, and it makes boring reading." https://news.ycombinator.com/newsguidelines.html
https://www.crunchbase.com/organization/modular-ai
Andrew Kelley predicts the future:
Also the team behind the project includes very smart people with a history of delivering other great projects (LLVM, Clang, MLIR and Swift).
I don't see how discouraging other people for trying to innovate is helpful. In the end this project could benefit all of us.
That's debatably where the trolling, misinformation, and other underhanded tactics come in. Maintaining the status quo and protecting one's interests, even to the detrimental of the whole, fairness, or the greater good.
> we expect that Mojo will be open-sourced. […] We don’t have an established plan for open-sourcing yet.
But I'm not trusting "don't be evil" tag lines anymore.
So we'll see when they actually make something open enough that we can entrust our programming stacks to it.
I think their strategy makes quite a bit of sense. Superficially Python like, can use Python libraries, fast and safe like Rust, automatically determines where algorithms should run (to some extent).
One thing that comes to mind however, is library performance. I'm assuming it will be ~3X faster than Python (like basic Python code) but if you are coming from Rust or C, this could be an issue.
My access was approved within one hour.
Lattner is an open source fan, I believe it will be open sourced eventually, but maybe the early Swift development was traumatizing. Just a guess.
(Actually IIRC I started learning it when they promised to open source it later that same year. So the actual moment of open-sourcing might have been slightly later.)
I meant when they actually opened the gates. My recollection is that they were overwhelmed by the feedback and community building.
I thought it was odd at first that they weren't letting me run it locally too, but if they control the compute environment then they can force push interpreter(?)/compiler updates and eliminate a lot of bug reports from folks that are just using an out of date build. Seems pretty reasonable to me for something that's so early in development.
There appears to be confusion here. Mojo is trying to align itself with and be a kind of superset of Python, like C++ was to C. It's also an alternative for Python users, to use it, instead of reaching for C++ or C when they need speed or going lower level. From their website, that's not what V devs ever claimed nor is it their angle for users. V is a different programming language (and compiled), which provides benefits such as greater speed, type safety, etc... in comparison to Python. It would be a switch over, not a superset.
V is more of an alternative to Go and then C (in terms of ease of use and greater safety). Mojo and V are mostly in different lanes, where they appeal to different audiences, and how they are marketing themselves has nothing to do with the other.
> Anyway, never gonna try it, unless I'm absolutely forced. I hate their marketing strategy, it sounds too bad
This comes across as weird, to claim one will never use or gives the impression to never have used the programming languages they are talking down on. If not using Mojo or V, why be so preoccupied? Just skip over to what one does like and not be worried.
> ...hate their marketing strategy
Just about all programming languages attempt to sell themselves to the public. If they don't, people don't hear about them, and they die in obscurity.
And a lot of times, people are allowing themselves to be brainwashed by what big corporations have to sell, because they have big budgets for marketing. People often like, what big companies told them to, because of their infomercials and marketing campaigns. Sometimes, it can be good or at least fair to give alternatives a shot.
On the front page of their website, they claim to have full compatibility with python code, an ownership model and a borrow checker, and being as fast or faster than C++. These are just some of the first things I noticed!
Apparently one of the people involved with the creation of this language have some credibility which is pretty confusing. They are the CEO of the company behind "mojo" as well.
I think this is going to be another V-lang situation personally, despite the credibility of Lattner. I think this is just someone seeing an opportunity to cash in on the "AI hype" situation that exists currently.
Even reputable people can be influenced to do shady things when lots of money is involved!
This is of course just my opinion and I wouldn't mind being proven wrong, but I'm definitely very skeptical right now!
Mojo is not only directed at AI, but at Python users as an alternative and superset. The use of it for AI, is partially based on the use case of Python. Insinuating that Mojo is trying to "cash in" on "AI hype", is odd in the context of all the others who clearly do so, but where is the "cash in" at? Mojo and Vlang are free.
Vlang (V-lang) has made no such marketing claims in regards to AI, so pairing the languages together comes across as odd. If the argument is about "cashing in", that looks even weirder. The users of both languages can get it for free. As with any programming language, there can be donations and sponsors. Why would it be a problem for Mojo or Vlang to have sponsors, but somehow not so for other competing languages, in regards to their claims? This comes across as at least one is fine with their favorite language getting sponsors or donations, but not others. Or maybe hoping to cut the funding sources of new languages, so there will be less competition.
> Even reputable people can be influenced to do shady things when lots of money is involved!
The accusation or insinuation is being made that Lattner is "shady" for money, while a backhand is being thrown at Vlang too. Outside of throwing smears, there is no proof of them being in it for hustling money or doing anything that different from what competing programming language creators have and are doing. And that's without talking about the corporate created languages, where obviously money and market share are major driving forces for them.