>>> 2024-07-13 the contemporary carphone (PDF)

Cathode Ray Dude, in one of his many excellent "Quick Start" videos, made an interesting observation that has stuck in my brain: sophisticated computer users, nerds if you will, have a tendency to completely ignore things that they know are worthless. He was referring to the second power button present on a remarkably large portion of laptops sold in the Windows Vista era. Indeed, I owned at least two of these laptops and never gave it any real consideration.

I think the phenomenon is broader. As consumers in general, we've gotten very good at completely disregarding things that don't offer us anything worthwhile, even when they want to be noticed. "Banner blindness" is a particularly acute form of this adaptation to capitalism. Our almost subconscious filtering of our perception to things that seem worth the intellectual effort allows a lot of ubiquitous features of products to fly under the radar. Buttons that we just never press, because sometime a decade ago we got the impression they were useless.

I haven't written for a bit because I've been doing a lot of traveling. Somewhere in the two thousand miles or so we covered, my husband gestured vaguely at the headliner of our car. "what is that button for?" He was referring to a button that I have a learned inability to perceive: the friendly blue "information" button, right next to the less friendly red "SOS" button. Most cars on the US market today have these buttons, and in Europe they're mandatory (well, at least the red one, but I suspect the value-add potential of the blue on is not one that most automakers would turn down). And there's a whole story behind them.

It all started in 1996 at General Motors. Wikipedia tells us that it actually started with a collaboration of General Motors (GM), Electronic Data Systems (EDS), and Hughes Electronics. That isn't incorrect, but misses the interesting point that EDS and Hughes were both subsidiaries of GM at the time. GM was a massive company, full of what you might call vim and vigor, and it happened to own both a major IT services firm (EDS) and a major communications technology company (Hughes). It was sort of inevitable that GM would try integrating these into some sort of sophisticated car-technology-communications platform. They went full-steam ahead on this ambitious project, and what they delivered is OnStar.

But first, a brief tangent into corporate history. I won't say much about GM because I am not an automotive history person at all, but I will say a bit about EDS and Hughes. Hughes was obviously the product of a notable and often enigmatic figure of history, Howard Hughes. GM owned Hughes Electronics because Howard Hughes had cleverly placed his business ventures under the ownership of a massive and brazen tax shelter called the Howard Hughes Medical Institute. Howard Hughes died without leaving a will or any successor as trustee of HHMI, putting HHMI into an awkward legal and organizational struggle as it abruptly pivoted from "Howard Hughes' personal tax scheme" to "independent foundation that incidentally owned a major defense contractor." HHMI ultimately made the decision to turn Hughes' business empire into an endowment, and sold Hughes Aircraft. General Motors was the high bidder. A set of confusing details of the Hughes amalgamation, like the fact that Hughes Aircraft didn't own all of the Hughes aircraft, lead to the whole thing becoming the Hughes Electronics subsidiary of GM.

After GM essentially stripped it for parts, Hughes Electronics lives on today under the name DirecTV. The satellite internet company that actually markets under the name Hughes is, oddly enough, one of the parts that GM stripped off. Hughes Communications became part of EchoStar, operator of Hughes Electronics competitor DISH Network, which then spun Hughes Communications out, and then bought it back again. You can't make this stuff up. The point is that the strange legacy of Howard Hughes, the HHMI, and GM's ownership of Hughes Aircraft mean that the name "Hughes" is now sort of randomly splashed across the satellite communications industry. It's sort of like how Martin Marietta still paves freeways in Colorado.

Electronic Data Systems is not quite as interesting, but it was run by two-time minor presidential candidate Ross Perot, so that's something. GM dumped EDS almost immediately after launching OnStar. EDS eventually became part of Hewlett Packard, which, by that time, had become a sort of retirement home for enterprise technology companies. It more or less survives today as part of various large companies that you've never heard of but have nonetheless secured 9-digit contracts to do ominous things for the Department of Defense.

What a crowd, huh? It's a good thing that nothing strange and terrible happened to General Motors in approximately 2008.

So anyway, OnStar. OnStar was, basically, a straightforward evolution of the carphone backed by a concierge-like telephone service center. In that light, it's an unsurprising development: the carphone was just on its way out in the mid-'90s, falling victim to increasingly portable handheld phones. Hughes, by its division Hughes Network Systems, was an established carphone manufacturer but seems to have had few or no offerings in the mobile phone space [1]. To Hughes long-timers, OnStar was probably an obvious way to preserve the popularity of carphones: build them into the car at the factory, with factory-quality finish.

GM had their own goals. Ironically, it is in large part due to GM's efforts that built-in telephony is so common (and yet so ignored) in cars today. The situation was much different in 1996: OnStar was a new offering, only available from GM. It had the promise of competitive differentiation from other automakers, but for that to work, GM would have to differentiate it from the carphones widely available on the aftermarket. This tension, the conflict between "we built a carphone in at the factory" and "carphones are going out of style," probably explains why OnStar marketing focused on safety and security.

"General Motors has come up with the ultimate safety system" lead a '96 newspaper article. Marketing materials prominently positioned roadside assistance, automatic emergency calls on airbag deployment, remote door unlock, and locating stolen cars. These were features that your average carphone couldn't offer, because they required closer integration with the vehicle itself. OnStar was more than a carphone, it was a telematics system.

"Telematics" is one of those broad, cross-discipline concepts that we don't really talk about any more because it's become so ubiquitous as to be uninteresting. Like Cybernetics, but without a tantalizing but lost historical promise in Chile. Telematics has often been more or less synonymous with "putting phones into cars," but is more broadly concerned with communications technology as it applies to moving vehicles. There is a particular emphasis on the vehicle part, and telematics has always been interested in vehicle-specific concerns like positioning, navigation, and the collection of real-time data.

Telematics was already a developed field by the '90s, although the high cost and large size of communications equipment made it less universal than it is today. OnStar would lead one of the biggest changes in the modern automotive industry: the extension of telematics from commercial and industrial equipment to consumer automobiles. In doing so, it would introduce select GM drivers to an impressive set of benefits, almost a form of ambient computing. It would also start a cascade of falling dominoes that lead, rather directly, to a remarkable lack of privacy in modern vehicles and getting an email that something mysterious is wrong with your car two to four hours after the tire pressure light comes on. The computer gives and takes.

And what of EDS? EDS provided the other half of OnStar's differentiation from a mere carphone. OnStar was not only integrated into your vehicle, it was backed by a team of Service Advisors with training and tools to use that integration. The OnStar equipment included a GPS receiver, still a fairly cutting-edge technology at the time, and continuously provided your location to the OnStar service center in Michigan. Advisors had access to maps and travel directories and the ability to dispatch tow trucks and emergency responders. They could even send a limited set of remote commands to OnStar vehicles. The infrastructure to support this modern telematic call center was built by EDS, and the staff of human advisors provided a friendly face and a level of flexibility that was difficult to achieve by automation alone.

Besides emergencies and roadside assistance, the advisors could solve one of the most formidable problems in automotive technology: navigation. When GM's advertising and press coverage strayed from emergency assistance, they focused on concierge-like services focused around navigation. OnStar could direct you to gas or food. They could not only reserve a hotel room, but get you to the hotel. If you have seen the wacky turn-by-turn navigation technology that proliferated in the late 20th century, you might wonder how exactly that worked. Did an advisor stay awkwardly on the line? No, of course not, that would be both awkward and costly. They read out driving directions, which the OnStar equipment recorded for playback.

I really wish I could find a complete description of the user experience, because I suspect it was bad. The basic idea of recording spoken guidance and playing it back for reference is a common feature in aviation radios, but that's mostly for dealing with characteristically terse and fast-talking ground controllers, and usually consists of a short playback buffer that always starts from the beginning. Given the technology available, I suspect the OnStar approach was similar, but just with a... longer playback buffer. Thinking about listening through the directions over and over again to find one turn gives me anxiety, but it was 1996.

Technology advanced like it always does, and by the mid '00s at least some GM vehicles had the ability to display turn-by-turn instructions, provided by OnStar, as the driver needed them. Fortunately there are videos from this era, so I know that the UX was... better than expected, but strange. It's odd to see an LCD-matrix radio display, with no promise of navigation features, start displaying large turn arrows and distances after an OnStar call. One of the interesting things about OnStar is that the "human in the loop" nature of OnStar features makes it sort of a transitional stage between cassette tapes and Apple CarPlay. OnStar allowed human operators and remote computer systems to do the hard parts, allowing cars to behave in a way that seemed very ahead of their time.

One of the interesting things about OnStar, given the constant mention of satellites in its marketing, was the lack of actual satellite communications. Hughes, a satellite technology company, was involved. Articles about OnStar coyly refer to satellite technology, or say it's "powered by satellites." Of course, OnStar cost $22.50 a month in 1996, and $22.50 a month didn't entitle you to so much as look at a satellite phone in 1996. The satellite technology was limited to the GPS receiver; all voice communications were cellular. AMPS, specifically. The first several generations of OnStar, into the early '00s, relied on AMPS.

Telematics, telemetry, and the applications we now call "IoT" often struggle with the realities of communications networks. AMPS, often just referred to as "analog," was the first cellular communications standard to reach widespread popularity. For over a decade, everything cellular used AMPS. Then CDMA and GSM and even, may we all shed a tear, iDEN took over. These were digital standards with improved capacity and capabilities. It was inevitable that they would replace AMPS, and with the short lifespan of a consumer cellular phone, devices without support for digital networks naturally faded away... except for a bunch of them. OnStar and burglar alarms are two famous AMPS-retirement scandals. The deactivation of AMPS networks in 2008 left cars and alarm communicators across the country unable to communicate, and prompted a series of replacement programs, lawsuits, trade-in deals, lawsuits, and more lawsuits that are influential on how cellular networks are retired today (meaning: as rarely as possible).

The obsolescence of OnStar equipment in older vehicles by AMPS retirement left a black mark on OnStar's history that still hangs over it today. It was, I think, a vanguard of the larger impacts of fast-changing technology being integrated into cars. While vehicles have indeed become more reliable over time, there is an ever-present anxiety that new cars are more like consumer electronics, built for a three-year replacement cycle. The forced retirement of half a million OnStar buttons is probably one of the most visible examples of automotive equipment failing due to industry change rather than age.

In 2022, 2G cellular service was retired in the United States. With it went another generation of OnStar-equipped vehicles. For a combination of reasons, though, both a more conservative approach to 2G retirement in the cellular industry and likely GM's planning further ahead, only two model years were impacted.

Incidentally, Ford also had an offering very much like OnStar, called RESCU and introduced in 1996 as well. It was pretty universally agreed by automotive journalists at the time that RESCU was more primitive than OnStar and amounted mostly to a knee-jerk "we also have one of those" response to GM's launch. RESCU is perhaps worth mentioning, though, for its contribution to the lineage of Ford's SYNC platform, at least in the form of gratuitous all-caps.

In 2002, GM offered OnStar for licensing to other automotive manufacturers. Subarus, among others, began to sprout blue buttons in the overhead. But what had happened to competitive differentiation? Well, automotive technology tends to go through two phases: First, it differentiates. Second, it's mandated. The originating manufacturer can make quite a bit of money off of both.

In 1995, a year prior to the launch of OnStar, the National Highway Transportation Safety Administration (NHTSA) was already investigating the possibility of an Automated Collision Notification (ACN) system. ACN would automatically call 911 in the event of a dangerous crash, improving driver safety. As far as I can tell, GM is not the origin of the ACN concept. NHTSA's work on ACN started with the National Automated Highway System (NAHS), an ambitious technology development program launched in 1991 that imagined a very different self-driving car from the ones that we see today. The NAHS involved mesh networking between automated vehicles to form "platoons," close-following cars (for fuel efficiency) that synchronized their control actions. The mesh network would extend to road-side signaling systems, and would lead eventually to the end of traffic signals as cars automatically negotiated intersection time slots.

The NAHS never came to be and probably never will, but the NHTSA's retro-futuristic graphics of '90s sedans linked by blue waves echo through my childhood like they do through the pages of Popular Mechanics and the academic literature on self-driving. Or, they did, until a new generation of Silicon Valley companies coopted self-driving for their own purposes. This is not an entirely fair take on the history, I am certainly applying rose-colored (or is it cerulean blue?) glasses to the NAHS, but I think it is hard to argue that there has not been a loss of ambition in our vision of the self-driving future. For one, we stopped drawing blue waves on everything.

Anyway, GM may not have created ACN. If anyone, I think that honor might fall on Johns Hopkins University. But they sure did get involved: by 1996, the year OnStar launched, Delco Electronics was building ACN prototypes for NHTSA. Delco Electronics, a division of GM (Delco's history is closely intertwined with that of Hughes in this period, parts of Delco were and would be parts of Hughes and vice-versa). Over the following years, GM really jumped in: OnStar was ACN, and ACN should be mandatory.

Here's the thing: it's never really worked. The move of introducing a technology and then pushing for it to become mandatory is a fairly well-known one in the automotive industry, and to its credit, has lead to numerous safety advancements in consumer cars and no doubt a meaningful reduction on fatalities (to its discredit, it is often cited as one of the reasons for steeply rising prices on new cars).

Universal OnStar has come tantalizingly close. Europe mandated "eCall," functionally identical to ACN, in 2018. I'm not sure how directly GM was involved, but there are GM patents in the licensing pool required to implement eCall, so it's at least more than "not at all." But despite its increasing presence, ACN isn't required in the US. Automakers aren't even consistent on whether it's standard or a paid add-on.

GM is still hacking away at this one... as recently as last year, GM was taking federal grants to study ACN and propose standards. In collaboration with CDC, GM developed a system called AACN that uses accelerometer data to predict the severity of injury to occupants and difficulty of rescue. It's installed in newer OnStar vehicles, and Ford has even licensed it for Ford SYNC, but the data rarely goes anywhere at all... 911 PSAPs that receive the calls from ACN systems aren't equipped to receive the extra metadata; extensions to E911 to facilitate AACN data exchange are another thing that GM is actively involved in.

GM really seems to have put a 30-year effort into mandating OnStar in the US, but they just can't get it over the finish line. In the mean time, OnStar has stopped mattering.

GM's program to license OnStar to other automakers was short-lived. I'm not sure exactly why, but GM also gave up on their "OnStar For My Vehicle" aftermarket product. Even as OnStar continued to gain features, its ambition waned. I think that the problem was simple: by the mid-2000s, putting a phone into a car was becoming pretty easy. Besides, the "connected car" offered too many advantages for any automaker to turn down. Can you imagine the benefits of storing location history for the entire fleet of vehicles you've sold? You can sell that to the insurance industry! You know GM did, of course they did, and of course it's the subject of an ongoing class-action lawsuit.

OnStar just stopped being special. I was actually a little surprised to notice that the blue button in the overhead of my modern Subaru isn't an OnStar button; Subaru stopped licensing OnStar in '06. It's just another manifestation of Subaru StarLink, a confusing menagerie of vaguely-telematic features that are mostly built on contract by Samsung. Once the car has an LTE modem for remote start and maintenance telemetry and selling your driving habits to LexisNexis, throwing in a button that makes a phone call is hardly an engineering achievement.

You know, sometimes it feels like smartphones can only incidentally make phone calls. With the move to VoLTE, it's not even really a deeply-embedded functionality any more. "Phone" is just an application on the thing that, for reasons of habit, we call a "phone."

The legacy of OnStar is much the same: of course your car can make phone calls, GM shoved a carphone in the trunk in 1996 and it's still in there somewhere. It's just one of a million things modern vehicle telematics do, and frankly, it's one of the least interesting ones. Ironically, GM is taking the carphone back out: in 2022, GM discontinued the OnStar telephone service. It's no longer possible to have a phone number assigned to your car and use OnStar for routine calls. Everyone uses an app on their phone for that.

[1] I am excluding here their satellite phones, although they were surprisingly advanced for the mid-'90s and probably would have competed well with cellular phones if the service wasn't so costly.