Going to prison for not doing your hw? And what does going to see a therapist have to do with anything? People with all kinds of problems go to a therapist.

Same! I am looking to learn a low-level language and want to learn Rust due to all the things I have been hearing. However all the backend stuff (python)for deep learning is written in C++and nothing yet for Rust. I would love to avoid C++ as the language seems to have a lot of quirks and bugs that take time to get used (from what I hear) and these days I like efficient languages.

As far I know there aren’t any good computer vision libraries for Rust?

Clearly, we are not the only ones feeling this way. I suspect that if PyTorch or TensorFlow were to provide support for rust as a first-class citizen they would get a whole new community of users very quickly -- people looking to write production AI/DL/ML code in a fast, practical, memory-safe, widely used language.

The only other "fast, practical, memory-safe, widely used language" that looks like a real alternative to Rust is Swift. But unlike Swift, rust was designed from the ground up for for data parallelism and other forms of concurrency with "zero-cost abstractions" the allow "fearless concurrency" (e.g., take a look at https://github.com/nikomatsakis/rayon/ to get an idea of what's possible), and for predictable performance (e.g., compiled code doesn't suddenly slow-down to collect garbage or synchronize reference counts among CPUs). Concurrency appears to have been an afterthought in Swift's case, explicitly: https://gist.github.com/lattner/31ed37682ef1576b16bca1432ea9... . See also https://doc.rust-lang.org/book/ch16-00-concurrency.html and https://blog.spencerkohan.com/impressions-of-rust-as-a-swift...

Data parallelism and other forms of concurrency should become more and more important over time as models grow larger and larger, well beyond GPT-3's 175 billion parameters. This would favor rust.

This is a great post. Time is limited and you have to make sacrifices somewhere.

Being sedentary seems to be really bad for you and programmers are the forefront of this as we spend lots of time motionless.

There seems to be correlations with major diseases as well as muscular skeleton pain. It’s not even 16 hours. Even 8 hours seems to be bad,(especially if it’s not spread out), since we live in a pretty sedentary world now.

I’ve been following these threads on hacker news as I want to prevent as much pain and diseases as I can.

I think sitting is more natural than standing. Hunter-gatheres sit around a lot but they don’t sit on chairs. On the other hand they don’t stand on one place for too long and usually are on the move.

https://www.sciencealert.com/hunter-gatherers-are-as-sedenta...

I was looking around for a language to write my own versions of convolution layers or LTSM or various other ideas I have. I thought I would have to learn c++ and CUDA, which from what I hear would take a lot of time. Is this difficult in Julia If I would go through some courses and learn the basics of Julia?

This would really give me some incentive to learn the language.

You could just use LoopVectorization on the CPU side. It's been shown to match well-tuned C++ BLAS implementations, for example with the pure Julia Gaius.jl (https://github.com/MasonProtter/Gaius.jl), so you can follow that as an example for getting BLAS-speed CPU side kernels. For the GPU side, there's CUDAnative.jl and KernelAbstractions.jl, and indeed benchmarks from NVIDIA show that it at least rivals directly writing CUDA (https://devblogs.nvidia.com/gpu-computing-julia-programming-...), so you won't be missing anything just by learning Julia and sticking to using just Julia for researching new kernel implementations.

In that benchmark, was Julia tested against CuDNN accelerated neural network CUDA code? If not, is it possible (and beneficial) to call CuDNN functions from Julia?

That wasn't a benchmark with CuDNN since it was a benchmark about writing such kernels. However, Julia libraries call into optimized kernels whenever they exist, and things like NNLib.jl (the backbone of Flux.jl) and KNet.jl expose operations like `conv` that dispatch CuArrays to automatically use CuDNN.

Yes! This is something I have noticed as well and it’s requiring a little rewiring of my brain.

Pretty cool. What did use to create it?(currently learning web development)

React and firebase :)

I’m primarily a windows user and am programming to create products. Since I don’t work a programmer job, my learning is unstructured. Is bash the main benefit of learning something like Ubuntu?

I have delayed making the switch because I have lot of things to learn (currently relearning math, testing out different deep learning architectures, and learning JavaScript (second language after Python)) and not sure if should prioritize learning it.

I still need to get better with GitHub which is something I clearly see the benefits of. Also need to get better with a text editor , using Visual Studio Code , mainly just for text completion in Python and JavaScript, so need to learn the additional benefits. Also want to test sublime text.

I see recommendations for vim, but I wonder if that will make me too twitchy and will start to type before thinking.

So due to the amount of things I am trying to learn, I have delayed learning Ubuntu.

I know Windows 10 now has built-in bash capabilities but I wonder if there are other benefits I am missing with Linux and worth learning.

> I’m primarily a windows user and am programming to create products. Since I don’t work a programmer job, my learning is unstructured. Is bash the main benefit of learning something like Ubuntu?

That's one advantage. The big advantage is that most enterprise web sites and webservices run either on a rented Linux server or on a Linux instance that's "serverless" (meaning it can run processes and access databases, but can't save stuff permanently to the filesystem). If you have familiarity with Linux, then that lends itself well to the work of troubleshooting the complications that come with that setup.

It's just easier to get going with programming stuff on linux (other than .net)

Need python? Ruby? docker? A different compiler or a specific library? They're just an apt-get install away. It just feels more natural and integrated.

Windows has always been an odd duck for programming on.

Thanks this is helpful to get started.