Laptops considered harmfulFriday, Dec 4, 2020 · 1800 words · approx 9 mins to read
Apple’s release of M1-powered Mac models and the introduction of the first line of Apple Silicon system-on-chip devices for the Mac line has me thinking about the future of personal computing. I write this on an almost fully loaded 2019 Mac Pro, bought mostly as a present to myself to celebrate a new job. Powered by one of the more powerful Xeon Ws in Intel’s current lineup, and configured in kind in the other major areas of its operation, the machine is an absolute beast.
Apple’s choice of Intel for this machine always felt a little strange in the light of AMD’s resurgent dominance in the very high performance workstation processor space, however it’s now crystal clear that Apple have a roadmap to beating the overall performance of even a machine configured like mine. They’re not there yet with the M1, and there’s clearly quite a long way to go — and neither will AMD and Intel stand still — but the writing is possibly on the wall. I figured I’d get about 5 years before they released a new Mac Pro that was faster than this one, which sits well with my personal upgrade cycle, but now I’m not so sure it’ll take that long.
M1 and its current performance-per-watt and performance-per-area, and the headroom they can grow into, with sensibly-sized chips that can be made with relative ease and relatively high yields, mean they likely won’t need that half decade. That’s also probably without being forced by the physics into a yield and scalability equation that’s only solved by chiplets. If they even need 3 years I’ll be surprised. It’s an incredible piece of semiconductor engineering from a world-class team, and the rest of the industry is now wondering how to compete. Competition is eminently possible, but just like in the smartphone and tablet space where they’ve long just run away with it, in the desktop computing space it’ll take some very special engineering indeed to stop the same thing happening there too.
For any customer that doesn’t actually care what operating system they’re using because they can all support whatever workflow and software the customer cares about, something that’s increasingly true in a Web-first world, many more eyes will be on the Mac product line. It’s time to assess where it now sits with the first M1-powered models, and where their product lines are likely to go in the future. I expect a significant continued growth in Mac business revenue and marketshare as a result, because most assessments will return favourable results.
That thinking about the M1 powered laptops in particular has also made me reassess a long-standing belief I’ve held for almost exactly 10 years now, as I stopped using a laptop as my main computing machine and switched back to desktops. My thesis is simple: portability is the only reason to choose a laptop. That might seem glib, dumb and obvious, but for many laptops are chosen even if they’re not the right type of computer to buy. If portability is the primary trait you need in your computer then great, but otherwise I think it’s always the worst choice due to the inherent compromises that must be made to bake portability into the system.
And for me, the compromises are vast. Small screens that, though very high quality these days at the upper end of the market, don’t lend themselves to good ergonomics that allow for comfortable use for any length of time without help getting them up off the desk somehow so that you’re not hunched over to peer at the cramped display.
Batteries that these days last long enough, but only when paired with low-power internals that are carefully designed, and judiciously driven by software, to sip from the battery rather than gulp. If your workload means you need to gulp then you’ll always find yourself near a power outlet, hampering the portability, and you’ll still be stuck using components that can’t lean too heavily on power draw because of thermals. Laptops, due to modern packaging, tend to have thermal systems with annoying fans that distract when they’re asked to work hard.
Apple have truly figured out the trackpad but their keyboards still leave a lot to be desired in my opinion, even with the improvements made in recent years to get away from the switch type that caused them all of their problems, and to improve the layout and usability. My strong personal preference is for keyboards with a lot more travel and tactility in their switches, and, crucially for me, some kind of rake to the presentation of the keys to the hand. Most laptop key beds are flat, which compromises ergonomics and usability for me. Most laptop keys also tend to be flat, which is an ergonomic compromise as well, since a sculpt, even just a reasonably shallow one, will help your fingers locate on the key top with less accidental slip off towards neighbouring keys.
Lastly we come to upgradability and expansion. I have a friend who makes a good living fixing and upgrading laptops because most modern ones are hostile to the user doing so themselves when it comes to the internals. Adhesives and solder conspire to take away options from the user when it comes to fixing or upgrading their machines, usually in the name of packaging to help with being thin and light, in order to afford better portability. And when it comes to adding functionality and expansion via the standard ports, that means visiting Dongletown and having to manage and bring those extras with you because they’re not built in, which perversely usually hampers portability.
All of that conspires to make me seriously averse to laptops as a general concept. The title of this post is designed to be playful rather than provocative, but it’s pretty much how I feel. I can only speak for myself of course, but computing on the go for me means my phone primarily, or my iPad at a push if I’m travelling. Laptops don’t work well on planes unless you’ve bought the most expensive seats, or have the most compromised sizes. A convertible or true tablet like an iPad does much better on a tray table.
Again, I’m not saying that a laptop isn’t truly a useful computing device for those where the portability is a hard requirement, especially given the relative restriction and compromise in the software and user experience on the best phones and tablet devices, but the second you sit down at a desk to use one you’re asking your body to do the wrong things in order to use it, without aids, and now you’ve traded the compromises that enable portability literally for your health.
Those aids — plugging it in to the wall; expanding it with dongles and adaptors to get more storage, a better webcam, or SD card readers; raising it up off the desk with a stand or using an external monitor so you’re not hunched over to see it; plugging in a proper keyboard and probably a mouse or external trackpad as well — just turn it back into a proper desktop computer and, for me personally, buying a proper desktop computer would always have been the better choice.
Then you don’t have to compromise as much on any of the display, performance and thermals, expansion and upgrades (and repair!), and ergonomics. Are those things worth the ability to move it around? For me that’s a resounding no, and so I’ve been a proper desktop computer user for a long time now because of that calculus. I’d rather be unable to use macOS on the move and use more appropriate — for me — devices on the go.
So what of the M1-powered laptops, then, and how they balance the equation. Well, they completely up-end it, don’t they, on the axis of performance and thermals? They do it in a such a remarkable way that I can’t really remember in the personal computing space, making them very hard to ignore even if portability isn’t a requirement. That you can now get serious desktop-rivalling performance in a very quiet package, because even the one with a fan never really gets that hot, portends the future of personal computing in a way I think many didn’t expect, and those that did more than they hoped.
There’s just a huge swathe of laptop and desktop machines now that are outclassed in one of the most restricted axes of compromise, and 2 of the 3 new machines are portable to boot. That means if you don’t need portability, Apple also stuck it in a really small, basically silent enclosure for you, which lets you set it to work at the heart of a proper desktop setup. If you can live with macOS, and most people could, you should be taking a serious look if your budget allows. It won’t be long before Apple have other system-on-chips for other models in their product line, presumably including the future variant of the enormous beast of a machine that I’m typing this on, that’ll similarly switch up what’s possible in a given performance and thermal envelope.
It ultimately doesn’t change the calculus for me personally, but I also don’t think we are that far away from a portable computer that truly could sit at the heart of a high-performance workstation setup and you’d get the best of all worlds if that’s what you want, with one of the areas of compromise basically wiped off the map. I’ll probably always prefer a machine that I can upgrade, expand and enjoy a higher performance and thermal envelope with, and so I’m really excited about what the future holds in that segment too.
So, yeah, I definitely mis-invested in this last hurrah of something Apple makes that’s powered by Intel, in the Mac Pro, but I love it to bits as a machine without any compromises, and I excitedly look forward to how quickly Apple’s own silicon can truly overtake its performance and what that’ll mean. Ultimately these new M1 machines allow for new workflows and possibilities in their respective segments that weren’t possible before, and that’s just at the low end of the consumer line-up. I can’t wait to see what’s in store in for the serious workstation end of their product line, and also what it means for the wider industry at large that now has to seriously catch up. That hammer blow of competition will hopefully mean exciting things in the non-Apple computing world as well as a result.