I personally work from home but go to a coffee shop as my “office” for a bit every day. Mobility is crucial to my, and many others’, workflows.
If you make 120k, travel once per month, and you waste ~3 hours of productivity each time you can't take your desktop with you... you would have made up for the difference between a $2k desktop and $4k laptop just in wasted labor.
If you work outside of those areas though, an Apple Arm CPU is amazing.
I was worried about software compatibility on m1 and was glad to have my intel imac for those things it couldn’t do, but in practice it now does everything. Some things require a bit of googling, but that’s par for the course for developer tools. And on the upside, the m1 runs rings around the 9600K in my iMac and it does this without a fan.
One of the reasons Mac was so nice was that everything worked the same locally as it did on Linux prod machines, whereas even today there are subtle bugs with naive Java or Python code that doesn't deal with paths or line endings correctly. On big multi-team projects, there are always some junior devs on Windows that just don't get it right.
This seems like another problem waiting to happen as more devs start using Arm, while the majority of developers and environments still use X86.
I don't push containers from my laptop to anywhere, instead those are built on CI/CD... and we are building both x86_64 and ARM containers these days as we are launching new workloads on Amazon's Graviton because it is cheaper and more cost effective.
The eco-system is not just x86 anymore.
For images with an Arm version available, and that's seemingly quite a lot, it mostly just works(TM). No Dockerfile changes needed.
If you have no other options, you can even run x64 stuff with a bit of low impact futzing. Docker Desktop will use QEMU to do this for you, nearly seamlessly. (though rather slowly)
The biggest gripe I have is the local volume perf, which is "tolerable." This is a problem for Intel Macs too, although to a lesser extent.
(edit for typo)
Possibly this is reflected in my general consumer use case, but all of my high computation oriented use cases are offloaded to a desktop.
Where are you deriving the majority of the value for this offering?
If I were to write tests, that is.
But at least now I know :D
That's very unusual. If you want a laptop for the easy stuff you're usually looking at a MacBook Air or iPad in the Apple ecosystem. These laptops are focused on people who want to be able to do "everything" on a portable computer.
At some point when you are "doing everything" you start to be tied to a physical location anyway. Multiple screens, hook ups to external equipment. So it just doesn't make much sense to me to actually need to do everything in a portable manner.
Edit: Apologies, I didn't address the part about the Mac OS ecosystem.
If you never require portability then a desktop is sufficient.
If you frequently require portability then it's worth getting an air and building your workflow (if possible) around offloading heavy computation to the network, e.g. heavy use of build servers, SSH, syncing.
But many people spend most of their time at a desk, but sometimes require portability. For example in a meeting, visiting a client, working from home occasionally, etc. For these people a heavy duty laptop is the best solution.
Like 3D rendering? Or deep learning training?
More broadly, that's why I put 'everything' in quotes. There will always be cutting edge workflows that require the power and cooling you can only get on a desktop or server, but that's a minority of a minority.
There are plenty of people who want to run multiple VMs, edit/render 4k+ videos, edit large audio projects, do semi-complex data analysis, or any of 100 other tasks without needing to choose power vs portability, or manage multiple devices.
1) Giving a bit less than half of it to the government and getting nothing, or 2) Giving all of it to Apple and getting a MacBook.
Works out to an effective 35+% discount on every business expense.
Or even if you are working on model that is on the server you still need to have good specs in order work with it.
Employers for their employees?
That used to be pretty common. But I'll almost certainly use one of these to replace ~2015 vintage iMac and MacBook Pro. Unless you want to keep the contents on a laptop and a desktop separate or have very high-end multimedia workload requirements, it probably doesn't make a lot of sense to pay for and manage two separate devices at this point.
I think a lot of pros these days don't.
From what I can tell this is very high demand. Obviously we didnt know enough about Apple’s launch planning and inventory but it seems like this only happens when there is very high demand.
Everything and more I need. Otherwise would need to fork up $20k for a mac pro
Regardless of whether GPU speed is a top priority or not.
Or is it merely competitive in specs or in applications like Illustrator?
The GPU performance of my M1 Mac Mini is roughly comparable to that of my Pixel 5 and Quest 2--both of which are running the Qualcomm Adreno 650 GPU.
I find it hard to believe and benchmarks [0][1] don't seem to support that.
[0] https://www.notebookcheck.net/Qualcomm-Adreno-650-GPU-Benchm...
[1] https://www.notebookcheck.net/Apple-M1-GPU-GPU-Benchmarks-an...
And they don't define "performance". Is it fill rate? Is it TFLOPs? Is it memory bandwidth? The M1 Max has 64 GPU cores. That sounds like a lot! It's not! The NVidia Laptop 3080 has 6,144 cores. That's NINETY SIX TIMES as many cores. I'm sorry, but supposedly achieving 90% of the "performance" of a GPU that has 2 orders of magnitude more cores is just not in the realm of believable. They are being intentionally evasive here.
https://images.anandtech.com/doci/17016/2021-10-18%2019_21_1...
The Max:
10.4 tflops
327 gtexels/sec
164 gpixels/sec