Feature Deterioration
2022-01-18
In 2018, Fabien Sanglard put into words what a lot of people around me had been expressing in one form or another - that increasing hardware power and internet bandwidth had not led to more functionality, but instead been used as a crutch for ever-more-awful software to do the same tasks less efficiently.
I want to linger on one specific aspect of this; that using technology used to be a constant effort to adopt new technologies - because they changed what was available to you. Graphics cards, processors, internet connections, monitors and peripherals in the 90s/2000s changed so rapidly that something made in 1995 was practically useless by the year 2001, because it simply was not powerful or featured enough to manage the new kinds of work you were doing in 2001. You adopted new technology to get more out of it. Today you adopt technology to not lose what you had.
For instance, in 2012 I had a phone with a replaceable battery, a 3.5mm jack, IR blasters, a removable SD card, and all-metal frame. In 2022 a Galaxy S21 has none of those things. Granted, it has NFC, BTLE, a 1440p screen, and three much-improved cameras. I like all those things. But i also liked the things we lost.
Moreover, phones today have no new features compared to phones five years ago (2017). Sure, there are more cameras, a physical zoom here or there, but none of this actually makes any difference. It’s not like you can’t get some pictures without them. Instead, we’re doing the same thing we were doing back then.
Limit of Improvement
The aforementioned S21 is 1440p, is that actually even perceptible? Sure, 800x600 looks like crap, but is a 2k or 4k phone really noticeably better than a 1080p one? I firmly fall in the "framerate over resolution" camp, but there’s a clear upper limit on both of those measures where further improvements simply aren’t noticeable - especially on a phone. Imagine we have 4k 165hz phone screens, isn’t that the limit of improvement for an LED display?
We’ve already reached this point with sound - nobody has sound cards or gives a second thought to sound reproduction - speakers capable of accurately reproducing any sound within hearing range and at high volumes are cheap and abundant. There are no techniques that excite anyone when it comes to a speaker playing sound - we don’t get animated by finding out that a phone can play >44khz samples, because we simply can’t hear it and it adds nothing.
Once we’ve reached that limit of improvement with video, we suddenly find that we no longer need vastly more storage - because the largest amount of data by far is audiovisual, and if we hit the LOI on those two things, the size of videos and pictures suddenly becomes fixed. They will not meaningfully increase in size over time. A 2hr movie will be N gigs, and there’s no reason for it to be larger; it offers nothing, and costs more space. Why would you want it?
Eventually, everything we do will settle at these "limits of improvement", and no improvement to specifications will provide a better experience. This is a bit of an uncomfortable proposition, because thusfar we’ve always had something that can be quantifiably better year over year. But what happens when there isn’t anything to look forward to? When cameras take pictures today just as well as they will in 10 years? When screens are as good today as they will ever be?
And, importantly, the vast increase in raw hardware power hasn’t translated to smoother experiences. Apps do not run better, faster, or without delay. All phones somehow manage to be slow to transition screens, press buttons, check notifications, or even just open an app activity. This can largely be blamed on apps stealing resources in the background, and increasingly-inefficient OS (specifically phone "desktop environment") features.
GG, 3g
All major telecomm companies in the US are shutting down 3g service this year. If your phone doesn’t have (or its software doesn’t support) Voice-over-LTE, then you’re back to 2g calls and texts. To say nothing of the cars, medical devices, cell modems, security systems, etc which will suddenly have swaths of features (if not the entire device) rendered useless.
This is something that occurs with software versions, with old ones losing support and new ones being required, but it’s the first time that it’s occurred with a cell network; something that hitherto has been indefinitely reliable.
You might say this is acceptable - after all, 2g is still supported, and that means dumb phones are still viable. But it closes the cheapest option for data transmission over cell networks, and poses the question of when (not if) 2g will go away. Or LTE. It suddenly seems as if not even cell networks can be relied upon.
Functionality Treadmill
I recently discussed with some friends the problems surrounding smartphones - addiction, distraction, loss of boredom, etc. And in that, a few expressed support for buying dumb phones to avoid all that. Instead, i put it this way;
Ive trimmed down the distractions over the years, its not bad now. I like the features I have. The problem is that its not possible to freeze that functionality in place; eventually updates will force upgrades, or degrade the experience so much its not worth it anymore Theres a constant effort to weedwhack new anti-features and find new solutions to replace old ones that worked great
This is ultimately my issue - I have what i want, but i can’t retain it. I have to adopt new devices just to retain the same functionality i previously had. My impulse to get a dumb phone is as a recognition that i’ll never be able to retain that same quality of experience that i have now - because i rely on software built by so many others who fundamentally do not give a shit about if their software is any good, or has any longevity.
If your new software doesn’t run on existing hardware, it’s worse than the old software.
Where are we going with this?
The crux of everything up to this point is to say that we must acknowledge that infinite improvement is impossible, and we’re fast approaching the best experience that anyone can have. When that occurs, we should (must?) focus on building more efficient and less bloated software, at least on client devices. A perpetual cycle of feature deterioration is not "sustainable", nor is it good for consumers. We cannot continue deprecating and degrading perfectly functional solutions after reaching the LOI for a feature.
Ultimately I hope that once we’ve reached these limits, companies begin to realize that new features don’t buy them anything, and instead aggressively offer devices meant for longevity - or at least easy maintenance.
