Well, I’ve worked for the government (as contractor), corporations, and small businesses, I could count a few times I’ve seen people using Apple Mac Pro devices on one hand (more often seeing Macbook Pro rather, but very rarely for development) and more time than I can count on either Linux or Windows workstation computers.
We use Linux desktop often, because most of our servers are running on Linux so it helps to have version conformity when matching up with server’s versioning and we occasionally use Windows for Visual Studio, proprietary software and so forth. But there are a few times where we get discounts for buying software for Linux rather than Windows.
Employees in my office switched from Apple Macbook Pro to Windows/Linux based laptops like Framework Laptop, because Macbook Pro often time lacked GPU that you would find on Linux and Windows workstation. Apple is going off on it’s own little world with their own Metal API/GPU and it doesn’t reflect the reality in real world emerging technologies. For instance, there are some computational challenges that in my office, we make use of Vulkan Compute so that we can purchase both Nvidia GPU and AMD GPU to generate real-time data, had we used Metal API and Apple’s products, it would’ve been cheaper to purchase cloud compute servers. (We wanted to ensure each developer can test the given Vulkan code on their own desktop/workstation.)
Yeah, and I am honestly surprised that you could do ok for AI on Mac since I was pretty sure that Tensorflow/Pytorch are pretty much CUDA implementation primarily and only have recently worked on branching out to other API.
My experience has been all GPU-intensive workflows have been pushed to the cloud. It works a lot better for CI/CD purposes as well, and most of the larger datasets are too practically large for your laptop, it ends up being prohibitively slow to download datasets from databases to your own laptop and then train on your local machine.
I could be biased since most of my network is in the startup scene in SV, where hardware cost is generally the LAST thing most companies worry about. I haven’t seen a non-mac software company that’s not a 5000+ dinosaur person company.
Disagree.
All the software companies i work with has switched to MacBook Pros as their mainline professional laptop of choice in the past decade.
It’s literally a better product for most of developer work and much easier to support.
In fact, I’m confident that MX MacBook Pros have cannibalized a good chunk of Mac Pro sales because they are just that good.
Well, I’ve worked for the government (as contractor), corporations, and small businesses, I could count a few times I’ve seen people using Apple Mac Pro devices on one hand (more often seeing Macbook Pro rather, but very rarely for development) and more time than I can count on either Linux or Windows workstation computers.
We use Linux desktop often, because most of our servers are running on Linux so it helps to have version conformity when matching up with server’s versioning and we occasionally use Windows for Visual Studio, proprietary software and so forth. But there are a few times where we get discounts for buying software for Linux rather than Windows.
Employees in my office switched from Apple Macbook Pro to Windows/Linux based laptops like Framework Laptop, because Macbook Pro often time lacked GPU that you would find on Linux and Windows workstation. Apple is going off on it’s own little world with their own Metal API/GPU and it doesn’t reflect the reality in real world emerging technologies. For instance, there are some computational challenges that in my office, we make use of Vulkan Compute so that we can purchase both Nvidia GPU and AMD GPU to generate real-time data, had we used Metal API and Apple’s products, it would’ve been cheaper to purchase cloud compute servers. (We wanted to ensure each developer can test the given Vulkan code on their own desktop/workstation.)
deleted by creator
Yeah, and I am honestly surprised that you could do ok for AI on Mac since I was pretty sure that Tensorflow/Pytorch are pretty much CUDA implementation primarily and only have recently worked on branching out to other API.
My experience has been all GPU-intensive workflows have been pushed to the cloud. It works a lot better for CI/CD purposes as well, and most of the larger datasets are too practically large for your laptop, it ends up being prohibitively slow to download datasets from databases to your own laptop and then train on your local machine.
I could be biased since most of my network is in the startup scene in SV, where hardware cost is generally the LAST thing most companies worry about. I haven’t seen a non-mac software company that’s not a 5000+ dinosaur person company.