Mojo – a new programming language for AI developers • Hacker News Board

Re: Mojo – a new programming language for AI developers

#251

Post by kaba0 » Thu, May 04, 2023, 2:54 PM UTC

ecs78 wrote: post #219

teleforce wrote: post #195

Earlier quoted context omitted.

Very strange to have no GC as a innovative feature for a modern programming language. Personally I think Dlang get it right by making GC as a default and provide no GC as an optional feature. As a comparison, auto industry is moving toward fully automatic transmission especially for the EV but software industry is still undicided and seems cannot even come up with a robust GC mechanism that is on par with no GC in te…

Traditional GC isn’t less complex to program then automatic reference counting. Traditional GC has its place in short running extension languages, but in longer running programs you run the same risk of memory leaks as automatic reference counting since you can still over-retain from a poor ownership model. What goes wrong is slightly different, but with automatic reference counting it is easier for the compiler to f…

RC is GC, but without cleaning up cyclic references. In general, leaking memory is not a security problem (and with a good debugger it is trivial to fix), so I don’t think it is too important a property. Leaking every cyclic reference is a bit more of a problem, but it has solutions like the occasional tracing phase like with a “traditional”, mark’n’sweep GC, but at that point why not just go with them as they are much more performant?

Re: Mojo – a new programming language for AI developers

#252

Post by kaba0 » Thu, May 04, 2023, 2:57 PM UTC

shrimp_emoji wrote: post #232

teleforce wrote: post #195

Earlier quoted context omitted.

Very strange to have no GC as a innovative feature for a modern programming language. Personally I think Dlang get it right by making GC as a default and provide no GC as an optional feature. As a comparison, auto industry is moving toward fully automatic transmission especially for the EV but software industry is still undicided and seems cannot even come up with a robust GC mechanism that is on par with no GC in te…

> Very strange to have no GC as a innovative feature for a modern programming language. It's because, first, there was manual memory management, which was error-prone. Then came garbage collection, which was safe but slow. Most recently came borrowing and move semantics, which offers the best of both worlds -- safety and speed -- at the cost of some (arguably justified) cognitive overhead and code flexibility.

GC is first and foremost a code design tool. Not having a GC causes less flexible code, every refactor will likely has to reimagine the memory ownership model at every relevant area.

So I absolutely don’t think of the borrow checker as superior to a GC, it’s a different tool with different tradeoffs. It was a good choice for a low-level language like Rust, but not necessarily for a high level language.

Re: Mojo – a new programming language for AI developers

#253

Post by kaba0 » Thu, May 04, 2023, 4:46 PM UTC

staticfloat wrote: post #16

cube2222 wrote: post #8

This looks quite interesting with big names behind it. I'd love to see a comparison vs Julia though, which I think tried to tackle some of the same problems.

It is a little disappointing that they're setting the bar against vanilla Python in their comparisons. While I'm sure they have put massive engineering effort into their ML compiler, the demos they showed of matmul are not that impressive in an absolute sense; with the analogous Julia code, making use of [LoopVectorization.jl]( https://github.com/JuliaSIMD/LoopVectorization.jl ) to automatically choose good defaults…

But that’s on different hardware than what they used.

Re: Mojo – a new programming language for AI developers

#254

Post by mkaic » Thu, May 04, 2023, 7:38 PM UTC

Archit3ch wrote: post #56

pknerd wrote: post #37

Earlier quoted context omitted.

I already added "Mojo Evangelism is my mojo" in my LinkedIn Title.

Dibs on Mojo Dojo (mojodojo.com), your new premier training resource for Mojo.

Mojo Dojo was briefly the name of a trampoline park in Sacramento before it got bought out by SkyZone

Re: Mojo – a new programming language for AI developers

#255

Post by throwawaymaths » Thu, May 04, 2023, 8:10 PM UTC

itissid wrote: post #243

throwawaymaths wrote: post #234

Earlier quoted context omitted.

> a new performant language that can call Python don't forget the control layer/data layer separation principle. Performance mostly only matters at the data layer, and I don't believe that python ML really has a substantial problem with this, aside from not having a real distribution story. So "having a more performant python" doesn't really solve that much. I'll tell you what could make the control layer better. - n…

> control layer/data layer separation principle Could you explain or give references to what you exactly mean by this? I've heard of separation of concerns, but is this a specific realization of that principle?

Ugh. I misremembered terms. Look up control plane and data plane. This is well understood.

Re: Mojo – a new programming language for AI developers

#256

Post by IceHegel » Thu, May 04, 2023, 10:03 PM UTC

earthboundkid wrote: post #72

One of the biggest problems with Python is packaging, and one of the biggest problems with Python packaging is installing scientific computing modules that rely on C/C++/Fortran. Will Mojo attempt to address these problems?

Yes this must be fixed. Python packaging is such a nightmare I use a fresh docker image for every project, which is much easier than messing with pyenvs, conda, broken wheels and that whole mess.

Re: Mojo – a new programming language for AI developers

#257

Post by tabtab » Thu, May 04, 2023, 10:08 PM UTC

ModernMech wrote: post #107

tabtab wrote: post #88

Is there a sample or tutorial that jumps right into typical examples of what's different or special about it? I tried to find such, but the docs get lost in syntax details and other minutia. I want the ADHD-meat-and-potatoes tutorial. Maybe that's asking too much, but I did it anyhow, as any ADHDer would.

> I want the ADHD-meat-and-potatoes tutorial. What does this mean?

"Gimme gimme gimme!"

Re: Mojo – a new programming language for AI developers

#258

Post by stallionz53 » Fri, May 05, 2023, 3:34 AM UTC

That video had really high production quality making it seem like Modular.com is very well funded. What is their source of funding, what’s going to be their business model?

Re: Mojo – a new programming language for AI developers

#259

Post by caleb-allen » Fri, May 05, 2023, 1:07 PM UTC

ChrisRackauckas wrote: post #31

chubot wrote: post #5

Seems like it could be interesting ... Initial impression is that the animation of Mojo code vs. Python code has a bad UX. Why not just show the code side-by-side instead of animating it and making me click? Another obvious question is how is it different than Numba and so forth? https://docs.modular.com/mojo/why-mojo.html The Mojo language has lofty goals - we want full compatibility with the Python ecosystem, we wo…

> A major goal of Clang was to be a “compatible replacement” for GCC, MSVC and other existing compilers. It is hard to make a direct comparison, but the complexity of the Clang problem appears to be an order of magnitude bigger than implementing a compatible replacement for Python. The journey there gives good confidence we can do this right for the Python community Is it though? For Mojo to be a compatible replaceme…

I saw your talk in London a few weeks ago and immediately thought of your discussions on python accelerators when I heard about Mojo - glad to see your thoughts here!

Re: Mojo – a new programming language for AI developers

#260

Post by ThorStone » Fri, May 05, 2023, 2:16 PM UTC

version_five wrote: post #68

newswasboring wrote: post #65

Earlier quoted context omitted.

Oh god I hope this is not a close source language. I have spent 10+ years fighting the MATLAB ecosystem, I prefer not to spend the next 10 years fighting this thing.

Matlab existed for a reason, and has since become irrelevant, in particular in the face of python being used as a more flexible open source scientific computing scripting language. If this is closed source, it's already as irrelevant as matlab so no reason to fight it. If there are useful bits there will be python versions of them.

Matlab is a lot faster than Python. At least an order of magnitude, once you move out of calls to LAPACK et al. It has been improved quite a bit on that front.

So Python is often not a realistic alternative. This is where initiatives like Mojo come in. They will be at least an order of magnitude faster than Matlab again.

But if you want to get people to move from Matlab to a powerful Open-Source alternative: Julia has a syntax that is much closer to Matlab than Python's. I had good success to get colleagues to use Julia which wouldn't look at Python, because the syntax was too far out of their comfort zone.