Why Dead Sonos Speakers Mean You’ll Never Own a Driverless Car
Over the past 15 years or so, Sonos has established itself as the purveyor of pricey speaker systems that come with a unique twist. The company developed a proprietary wireless communication system that enabled multiple speakers to sync up for easy whole-home audio. But last month, Sonos announced that it will end support for many of its oldest wireless speakers. This is just the latest in a recent string of such moves by technology companies.
Over the past 40 years, as technology has permeated every aspect of our lives, many of us have become accustomed to a different pattern. For many of the things we use like computers, smartphones, or even speakers, the product itself can actually evolve and maybe even improve over the course of its lifecycle. If you use a computer, operating systems get periodic updates with new features and capabilities. The same is true of most mobile devices.
By the time the initial battery will no longer hold a charge, the core functionality of that device may have gone through multiple iterations. One of the best examples of this process is the Apple Watch. When it debuted in 2015, the software running on it had some interesting features, but much of it simply wasn’t that usable. Each year since, Apple has rethought the interface and by 2019, those first-generation watches were fundamentally different products from when they were first purchased. They had also reached the end of life for support and would not get any further upgrades.
The beauty of software-defined systems like computers, mobile devices, and smart speakers is that new software can transform a product without touching the hardware. Up to a point. There are of course limits beyond which older hardware can no longer support new features. That’s why early generations of the Apple Watch can’t run WatchOS 6. Sonos speakers that were introduced earlier than 2009 won’t be able to get the same features as the latest Sonos Move speakers. It’s also why late-1990s GM cars with the connected OnStar service stopped working after analog cell phone networks were decommissioned.
But it’s not just technical hurdles at play here. These are businesses, and for a business to survive, it needs to bring in more revenue than it spends. The salaries of the engineers who develop this new software cost money. The infrastructure to deploy this software costs money. Once a customer has paid for their watch or speaker, the manufacturers aren’t getting any incremental revenue unless customers buy additional products.
At some point in the life of every product there comes a time when manufacturers have to draw the line and say, I’m not going to support this any more. The alternative is simply economically untenable.
Why the tech inside a Tesla changes everything
Until the rise of Tesla, end-of-support woes have never really been a problem for the auto industry, where support traditionally ended when you drove the car off the lot. This changes radically with the arrival of automated vehicles, and it’s why you will probably never own one.
For more than 130 years, the automotive business model has been to design, build, and sell vehicles to customers (or more precisely to dealers, who then sold to customers). Once you had the title to that vehicle, you were largely on your own. For some number of years after purchase, manufacturing defects would be fixed under warranty. But aside from the inevitable effects of wear and tear from actually driving, the vehicle stayed fundamentally the same. Sure you could buy aftermarket upgrades and modifications, but by the time the car was in your garage, the engineers who created it were already working on the next generation.
But Tesla brought the software support philosophy to the auto industry by making its vehicles capable of accepting over-the-air updates. The company has regularly deployed software updates that add new functions, improve driving range, or even make the car accelerate faster, mostly at no additional charge to the customer.
Compared to a car, even pricey Sonos speakers or MacBook Pros are relatively modest in price. Up to this point, Teslas have been by far the most software-defined vehicles ever produced. A customer can pay to add the AutoPilot driving assist features to a three-year-old car with an in-app purchase and OTA update because the hardware to accept those updates was built-in at the factory. But you can’t do that to a 2014 model, because the necessary sensors and actuators just weren’t there.
Software crashes, random reboots, and flaky sensors are simply not acceptable.
For much of the past decade, we’ve been hearing that the automated (or autonomous or self-driving) car was just around the corner. These vehicles will begin trickling out onto our roads in the next few years. Setting aside the question of the quality of Tesla’s “Full Self-Driving” option, which is a subject for a whole other rant, almost none of us will ever own a highly automated vehicle. The reasons why have at least something to do with the uproar over Sonos declaring that some of its older products would no longer be able to provide their existing capabilities alongside newer versions.
Automated driving in a vehicle is a safety-critical function. If your Sonos speaker dies, you might be annoyed and have to go dig some Bluetooth speaker out of a junk drawer. But the threshold for a minimum viable product in automated driving is orders of magnitude higher. Software crashes, random reboots, and flaky sensors are simply not acceptable.
While ongoing software support and updates are a nice to have for your infotainment system, they will be essential for AVs. These vehicles are also going to require regular servicing, such as testing and calibration of sensors. Today, when a car owner needs to replace an out of warranty part on their car, they are probably more likely to go to a discount auto parts store or garage and have the service done with third-party generic parts at a fraction of the cost.
When Sonos announced the end of support for older devices, they recommended disconnecting them from the user’s network, warning that not doing so might prevent newer devices from getting updates that are still available (for now). There have been plenty of valid complaints from customers who bought devices and want them to just keep working as they have. In the AV space, vehicles that don’t stay up to date and within specifications could pose a serious risk not just to the direct users, but to others on the road, including cyclists and pedestrians.
When Cruise, the AV development company owned by GM and Honda, unveiled its first purpose-built vehicle recently in San Francisco, called the Origin, founder and CTO Kyle Vogt discussed the design of the machine. Traditional vehicles are typically designed for a useful life of 10 years and 150,000 miles, although most last far longer than that. The average age of vehicles on the road in the U.S. today is 12 years and 25 to 30 year old cars are not uncommon (including the car in my garage that was built in October 1989). Except for wear on parts like tires and wiper blades, all major systems are designed for a similar lifespan.
The Origin has a modular design because it is designed for a 1 million mile lifespan. As a shared mobility vehicle, it is expected to have much higher utilization than individually-owned vehicles that typically sit parked 95% of the time. These vehicles may accumulate 100,000 miles a year or more. A traditional vehicle would be worn out in two to three years at this rate. The structure, chassis, and electric propulsion of the Origin are expected to last 1 million miles.
But systems like computers, sensors, and even much of the interior of the car are intended to be easily replaced every two to three years as the technology evolves and they become obsolete. If an individual were to own an autonomous car like this, much of the hardware would be way out of date before it stops functioning. Asking customers to replace such components more frequently would be cost-prohibitive, but not doing so would potentially make the vehicle unsafe to use.
The software also needs continuous updating, including the HD maps, security capabilities, and the functional and safety aspects. Again, asking customers to directly foot the bill would be too expensive. But integrating all of these costs into the fare for an automated ride-hailing service — the kind of service that would use the Origin — allows the costs to be spread among more users.
Continuing down the personal ownership route for AVs is probably an untenable proposition. Much like Sonos, Apple, and other tech companies, the manufacturers of these vehicles will not be in a position to support those customers indefinitely but unsupported hardware can’t realistically be left on the road for safety reasons. A recent blog post by Cruise CEO Dan Ammann was titled “We Need to Move Beyond the Car.” While the title is an oversimplification since it’s hard to argue that vehicles, like the Origin, are not cars, the century-old culture of how we use cars is definitely going to change.