Smart Devices Will Eventually Die, and the Internet Is to Blame

We may have made a horrible mistake by unnecessarily making our consumer electronics devices smart—and removing generations of future use in the process.
March 10, 2020, 12:00pm
The first-generation Sonos PLAY:5, one of the devices that the smart speaker company announced it was recently retiring from updates. Image: Sonos

A version of this post originally appeared on Tedium, a twice-weekly newsletter that hunts for the end of the long tail.

Recently, I had a bizarre customer service experience involving a company that I have been critical of at many junctures over the years. That company was Apple, and the reason I was on the phone with them involved my wife’s smartphone of choice, the iPhone 5.

Sold to her in late 2012, the phone has somehow managed to avoid device upgrades for nearly eight years (despite my semi-frequent pleas that she upgrade to a new device). But when we had problems logging it into iCloud recently, I chose to call Apple, and shockingly found them not only willing to service this device over the phone, but work through a variety of solutions on the phone for many hours as we attempted to figure out why it wasn’t logging in.

It took a few calls and some in-depth diagnostic work, but we figured out the problem, and earlier this week I was able to get the phone working again, with Jesse’s help. We kept a vintage smartphone out of the waste bin for another day.

It was not what I was expecting—it flew in the face about what I know about Apple and upgrades. But the fact we’re seeing problems in the first place reflects something that has been on my mind a while: The lack of consideration towards the upgrade and decay cycle in modern tech, particularly in terms of consumer goods, is going to bite—hard—in a few years. And not every company will be as understanding as this Apple support tech person (his name is Jesse) that clearly deserves a raise.

It’s time to have a talk about the coming gadget apocalypse that we haven’t been preparing for. Strap in.

Moore’s Flaw: Part of the problem we’re seeing is a computing mindset brought to standard consumer electronics

Recently, the speaker company Sonos has been taking a beating in the press for revealing an inevitable, but likely avoidable fact: Its early devices won’t last forever.

Founded in 2002, Sonos is one of the first companies that found success taking a traditional piece of electronics, the speaker, and making it “smart.”

The company, in its efforts to encourage those customers to upgrade and discourage the use of old speakers or bridge devices on its cloud servers past the point of old age, effectively kneecapped them—they can’t be upgraded to next-generation software, and keeping them in your setup could prevent your other devices from getting upgrades, too.

Sound like a bum deal? Fortunately for their investors, they sell a replacement.

The company attempted to clear up its self-inflicted mess by pointing out the devices will still work even without updates, but even with the clarifications, it’s still sticking with the original plan—no updates after May.

In one sense, you can’t blame Sonos for not planning for a future like this. As a startup, how was it supposed to know that it would be supporting speakers that it sold more than a decade ago? The odds were even it was setting itself up for an acquisition, a long life, or a noble failure.

But on the other hand, we have different expectations for audio equipment than we do computing devices. Think about it this way: If you buy a 1960s-era Fender Stratocaster from a pawn shop and it’s still in relatively good condition, and plug it into an amplifier that was produced this year, that guitar is still going to work.

This Bose 901 speaker will continue to work decades from now. Image: Automotive Rhythms/Flickr

If you plug a pair of vintage Bose 901 speakers into a modern audio system, those speakers should work, because ultimately the principles on which Dr. Amar Bose built that equipment haven’t changed in 50 years. It’s what allows a MiniDisc player, a record player, and an iPod dock to live on the same bookshelf.

The problem is, computers have never had these kinds of expectations around them. Sure, you can find some legacy ports on modern computing devices—this recent Vaio laptop has a VGA port, which made its first appearance on a computer in 1987 and still has use cases in boardrooms where old projector equipment lives—but for the most part computing equipment doesn’t work like that. It goes through generational phases.

Part of the reason for this is that computers move faster than other kinds of electronics. While some of the concepts, like soldered integrated circuits and processor sockets, mostly work the same generation after generation, the technology simply moves too fast to allow for hand-me-downs. If you were to reuse a desktop computer from 1996 and upgrade it to account for modern needs, basically the only thing you might be able to reuse is the case, which likely utilizes the ATX form factor.

Planned obsolescence has been built into the computer model since the beginning. In many ways, Moore’s Law, Gordon Moore’s observation that computing power would keep improving exponentially, detected the trend early and has at times enabled it.

Applying these standards to electronics that could once last generations creates a whole lot of discomfort. This can be seen, for example, in the case of the Apple Watch. Recently, Apple’s watch line, a somewhat minor part of its financial picture, outdid the entirety of the Swiss watch industry, which has been around for hundreds of years and has a completely different value proposition than your average laptop.

If you sell gadgets more often, it’s easier to print money. But it means that devices that were once built to last are now suddenly targets of planned obsolescence.

Look, technology moves fast. For decades, the world of computers surged past gate after gadgetary gate in search of the next new thing. The problem with Moore’s Law is not that it didn’t encourage more innovation, but it didn’t account for what we did with the old innovation.

Now, add the internet to the mix, and make it a defining element of its use case. And the problem becomes obvious. Companies don’t want to have to think about things they sold 20 years ago, but the smart device model, by default, requires that they do. Or, it should.

Let’s take another look at Sonos here. This is a company that leapt into a space where speakers could last generations, and decided that it couldn’t even let its devices survive for a single generation. Sure, it will still work, the CEO says, but the lack of updates clearly tips the scales in favor of a future upgrade. Computing power had advanced too much for those old workhorses to stay in use.

You can sell a 30-year-old car and it maintains some semblance of value, especially if it’s been well-maintained. Products like baseball cards and books still remain things people want to buy many years after they were first produced. But if a smart speaker company can’t promise that your internet-enabled device will be able to hook up to the internet eight years after you bought it, it’s useless.

Because, remember, a loss of updates doesn’t just mean you’ll get the fanciest new features—but it means you won’t get access to security updates that will keep the device alive for decades to come. Look at what I’ve already seen from my Mac Mini! Many consumer products have far higher standards for longevity than 15 years.

By allowing computers to infiltrate everything else—by adding things to our internet—we’ve decimated the long-term value of these products unnecessarily, all for someone else’s short-term gain.

And this is only the first wave of a problem that’s going to get a hell of a lot worse before it gets better.

Early-generation Roku devices, like the Roku XDS, recently lost access to Netflix. Image: Roku

The Sonos speaker saga is the first wave of what is likely going to be an entire generation of stuff broken by the internet

As I mentioned earlier, Sonos was early to this concept—not as early as, perhaps, The Clapper, but early enough that it might have been the first internet-connected smart device that most people might have encountered.

Another company that was early to this smart device model, Roku, also retired some of its early models recently, and Netflix stopped working on some of Roku’s early-gen devices, along with some smart TVs of the period made by other companies.

As I wrote a few weeks ago, it doesn’t feel quite as bad in the case of Roku, because they charge so little for their devices that buying a replacement box is trivial and worth the cost.

But the news about the early-gen smart TVs also getting the boot gives me pause. These sets, while no longer the latest and greatest and largely predating the recent 4K trend, are perfectly fine televisions. They should work for the next 20 years without a problem. But because Netflix arbitrarily raised its standards, it stopped supporting these sets, which is annoying.

While there is a way to work around losing Netflix on a smart TV set—easy, buy another set-top box that supports the device in some way—it just feels wasteful and cumbersome. The functionality is already in the set, after all! Instead of letting these sets slowly lose functionality, we should offer simple hardware upgrades that keep them up to date with modern standards while not limiting the other 95 percent of an otherwise perfectly fine TV. Put a slot in the back that upgrades the device’s brain. Easy.

These devices came out in 2010 and 2011 or so, and what worries me is what is to come.

In the decade between the release of the first smart TVs and now, the smart device trend really picked up in earnest, affecting things as varied as watches, drinking cups, thermostats, smoke detectors, toothbrushes, and even smart Gibson guitars (which didn’t work out for the guitar company).

Already, new signs are emerging. Just last week, word surfaced that Philips was going to stop updating early Hue Bridge devices, limiting their future connectivity and threatening their security.

Image: David Berkowitz/Flickr

According to a 2009 study from the National Association of Home Builders and Bank of America, the average thermostat is supposed to last 35 years. Can we trust that Google is going to support first-gen Nest thermostats for that long? It’s not like a smartphone.

Extend this to every device that you own that has a computer in its brain and a connection to the internet, that didn’t have those two things a decade ago, and you see the problem.

The microchip, in the long run, has turned things that were functionally fine (if “dumb”) into devices that may not make it into a second decade if they fail to get the ongoing support they need. Many of these devices were built by startups that have since left this world; others have become the victims of lacking warranties. Sure, in some of these cases, you can work around the faults of these things, but smart devices give the sheen of planned obsolescence to objects that could have lasted decades without continued internet access.

That awkward conversation Sonos just had with its customers about smart devices? Expect lots of other companies—including big ones—to have similar discussions in the coming years, with little to no path to repair in the future.

In the span of a single decade, we basically let a computer-centric mindset around planned obsolescence threaten to ruin the long-term usability of entire categories of products unnecessarily.

The solution here is not to fret or just drink water out of regular non-smart glasses, but to push electronics and gadget manufacturers to do better. If a device seems like something that should not stop working after eight or nine years, they need to guarantee upgrades more than a decade in the future. The design of these smart products, when applicable, should allow for user-replaceable hardware.

And for devices where their long-term use must be guaranteed, they need to offer dumb versions with the capability of getting smart upgrades. Let the user decide if they want a computer in their oven—and make that computer easy to plug in and replace, so that a decade down the road, they’re not stuck with a failed investment of a device.

We can’t bank on large devices being smart forever. We should have the ability to remove that functionality—or upgrade it as needed. That’s how we ensure hand-me-downs are worth handing down decades from now.

If they can’t promise that, don’t go smart.

Going back to my recent phone call with an Apple customer support rep, I think that the superhuman effort to attempt to get this phone working past its expiration date was super-noble.

But on the other hand, I wonder why it should have been super-noble. (No fault of Jesse!)

The base of the problem we’re seeing with this not-particularly-vintage gadget comes down to space and lack of upgrades. Apple has never offered a device with a MicroSD card, but for years sold devices with so little space that you would suddenly be out of storage with less-than-normal use. This is bad enough on its laptops, but in the case of its phones, it feels increasingly unforgivable over time, because it kneecaps these devices unnecessarily. I mean, this whole saga began because her 16-gigabyte phone runs out of space basically daily, and I tried to delete and redownload some apps for her to clear the phone’s cached space, and found what appeared to be aggressive bugs in iCloud in the process.

This phone is nearly eight years old and hasn’t received software updates from Apple for roughly three years. Image: Lucy Takakura/Flickr

I’ve been trying to figure out why Jesse cared so much about this phone when everything else about Apple’s actions has traditionally suggested that they would leave my wife’s phone high and dry. The best answer I have is that Apple, while I don’t think they’ve learned their lesson, is realizing this approach is not sustainable in the long term—at least not with these specific gadgets. Or at least some of their employees are.

I didn’t call and say, “My wife’s phone isn’t working.” I called and said, “My wife can’t log into her iCloud on her phone.” That may seem like a small difference, but I think it’s a significant one in terms of how Apple responded to this. It’s the difference between, “I’m having trouble with something you no longer support,” and, “This problem could cost you money now if you don’t fix.”

Technology is going to keep evolving and over time we are going to reach theoretical limits of devices. But if they’re talking to the cloud and people are paying money to access those cloud services, eventually those devices are going to matter less to our tech companies. The cloud is going to matter far more—and it’s going to be in their interest to keep these devices functional in the long haul because their profitability will continue to matter even as the device ages.

I can’t imagine that Netflix will unceremoniously kill another generation of smart TVs without thinking really hard about it—because killing the next gen of smart TVs will hurt a lot more than killing the last one. It will cost them money.

I hope the pendulum shifts in this way. Because that is how planned obsolescence will lose.