The first self-driving car fatality proves nothing

This morning’s Observer column:

In the US, about 33,000 people are killed in automobile accidents every year. That’s 90 a day on average. So on 7 May, about 89 other people as well as Joshua Brown were killed in car crashes. But we heard nothing about those 89 personal and family tragedies: the only death that most people in the US heard about was Mr Brown’s.

Societies have to decide what they want to do about automobile safety. It will come down to a cost-benefit analysis
Why? Because he was driving (or perhaps not driving) a semi-autonomous vehicle. Writing from Detroit (coincidentally, the capital of the traditional gas-guzzling, emission-spewing automobile), two New York Times reporters wrote that “the race by automakers and technology firms to develop self-driving cars has been fuelled by the belief that computers can operate a vehicle more safely than human drivers. But that view is now in question after the revelation on Thursday that the driver of a Tesla Model S electric sedan was killed in an accident when the car was in self-driving mode.”

Really? With whom is the safety of self-driving cars in question? Not with anyone who knows the facts about the dangers of automobiles…

Read on

The ‘Internet of Things’ will need better things

This morning’s Observer column:

You know the problem. You’re going abroad for a couple of weeks, during which time your house will be empty. You haven’t yet got round to installing a burglar alarm. But not to worry – just pop round to a supermarket and buy a couple of timer sockets. Plug them into the mains, set the timers to switch on and off at appropriate times twice a day, plug your lamps into them and off you go. Easy, peasy!

Well, yes. But it’s so 1950s. So analogue. Why not be really cool and have a proper networked timer socket, something that you can control from your smartphone from anywhere in the world? Something like the AuYou Wi-Fi Switch for example. Looks like it’s just the ticket. Plug it in, hold down the power button and it hooks up with the app on your (Android) smartphone, and – bingo! – job done. Now, where did you put that boarding pass?

But, hang on. Maybe you should just check the product reviews, just to be sure. Ah, here’s one by a guy called Matthew Garrett. “There’s a lot to like about this hardware,” Matthew writes, “but unfortunately it’s entirely overwhelmed by everything there is to hate about it.”

Eh? Turns out that Mr Garrett knows a lot about computer security…

Read on

The Apple watch: a solution in search of a problem

As some readers know, I’m not a fan of the Apple watch. Nor, it turns out, is the Guardian‘s Alex Hern:

The future of the watch can’t be the same iterative improvements that Apple has pulled off with the iPhone, iPod and iPad. The interface is just too ill-thought-through to work, even if the device itself is sped up significantly. But the most obvious alternative is to massively increase the amount of voice control the watch offers, and Apple simply doesn’t have the technical chops to do so. While Google and Amazon have been creating voice assistants that people seem to actually use and wax lyrical about, Apple … hasn’t. There’s no easy solution there.

But the saving grace for Apple is that the broader problem isn’t the company’s fault. It’s that smartwatches are a solution in search of a problem. A technology created, not to serve consumer demand, but to serve the need of device manufacturers to fill the revenue hole created by declining smartphone growth. You don’t need one, and neither do I. It just took me nine months of wearing it to realise.

Yep.

BUT… This thoughtful comment from a reader:

If you have a significant medical condition like Diabetics, particularly if you have diabetic kids, the Apple Watch and their Android equivalent is revolutionizing how you can go out in the world. I have neighbors who would have never allowed their 8 year old daughter go out for a sleepover because they need to check their blood sugar levels so often and now all they have to do is glance at their watch to track their behavior. Previously that sort of equipment was over $50,000 and now you can do it for a few hundred.

The spy in the cab

Interesting news for Tesla owners:

Everyone makes mistakes, and many people try to cover them up. But if you try to hide an error made behind the wheel of a car made by Tesla Motors, you are liable to be caught out. In fact, trying to hide what really happened in any kind of car accident could soon become just about impossible.

That’s the lesson of an incident over the weekend in which the owner of a Tesla Model X SUV crashed into a building and claimed it had suddenly accelerated on its own. But Tesla vehicles are constantly connected to their manufacturer via the Internet, and the company had this to say in a statement to the Verge:

“Data shows that the vehicle was traveling at 6 mph when the accelerator pedal was abruptly increased to 100 percent … Consistent with the driver’s actions, the vehicle applied torque and accelerated as instructed.”

Quite so. So what happens when your Tesla has been hacked and the hacker instructs it to accelerate into a wall? Or into a crowd?

Why the arrival, not the journey, matters

I have an article on the evolution of the Internet in a new journal — the Journal of Cyber Policy. I was asked to give a talk at the launch last week in Chatham House, home of the Royal Institute of International Affairs in London. Here’s my text.


One of my favourite autobiographies is that of Leonard Woolf, the saintly husband of Virginia. It’s a multi-volume work, but my favourite one is the volume covering the years 1939-1969. It’s entitled The Journey, Not the Arrival, Matters and it came to mind when I was pondering this talk, because in the case of the evolution of digital technology I think it’s the other way round: the arrival, not the journey, matters. And I’d like to explain why.

In 1999, Andy Grove, then the Chief Executive of the chip manufacturer Intel, said something interesting. “In five years’s time”, he declared, “companies that aren’t Internet companies won’t be companies at all”. He was speaking at the peak of the first Internet boom, when irrational exuberance ruled the world, but even so many people though he was nuts. Was the CEO of Intel really saying that all companies needed to be selling information goods by 2004?

In fact, Grove was being characteristically perceptive. What he understood — way back in 1999 — was that the Internet was on its way to becoming a General Purpose Technology or GPT, like mains electricity, and that every organisation in the world would have to adapt to that reality. So on the big story, Andy was right; he was just a bit optimistic on the timing front.

My article in the first issue of the new journal is entitled “The evolution of the Internet”, but the real meat is in the subtitle: “From military experiment to General Purpose Technology”. I say that because as the network has been evolving we have focussed too much on one aspect of its development and impact — namely the production, consumption and exchange of information goods — and too little on the direction of travel, which — as my subtitle implies — is towards becoming a GPT.

Arthur C Clarke is famous for saying that any sufficiently advanced technology is indistinguishable from magic, and for most of its users the Internet already meets that requirement. As Eric Schmidt, Google’s Chairman, once observed, it is the first technology that humans have built that humans do not understand. But while a General Purpose Technology may or may not be incomprehensible to humans, it has impacts which are visible to everyone

This is because GPTs have an impact on the world way beyond the domain in which they first appeared. They are technologies that can affect an entire economy and “have the potential to drastically alter societies through their impact on pre-existing economic and social structures”. Think steam engine, electricity, electronics, the automobile. GPTs have “the potential to reshape the economy and boost productivity across all sectors and industries, like electricity or the automobile”. And these transformations are about far more than simple technical innovation, because they often require the wholesale remaking of infrastructure, business models, and cultural norms. GPTs are the motive forces behind Joseph Schumpeter’s waves of ‘creative destruction’ and in that sense leave almost nothing untouched.

But if, as now seems obvious, the Internet is a GPT, then our societies are only at the beginning of a journey of adaptation, not the end. And this may surprise some people because the Internet is actually rather old technology. How you compute its age depends really on where you define its origins. But if you think — as I do — that it starts with Paul Baran’s concept of a packet-switched mesh in the early 1960s, then it’s now in its mid-fifties.

So you’d have thought that our society would have figured out the significance of the network by now. Sadly, not. And that’s not because we’re short of information and data about it. On the contrary, we are awash with the stuff. Our problem is that we don’t, as a culture, seem to understand it. We remain in that blissful state that Manuel Castells calls “informed Bewilderment”. So a powerful force is loose in our societies and we don’t really understand it. Why is that?

One good reason is that digital technology is incomprehensible to ordinary human beings. In that sense, it’s very different from some GPTs of the past. You didn’t have to be a rocket scientist to understand steam power, for example. You might not know much about Boyle’s Law, but you could readily appreciate that steam could powerfully augment animal muscle power and dramatically speed up travel. But most people have very little idea of what digital technology can — and potentially could — do. And this is getting worse, not better, as encryption, machine-learning and other arcane technologies become commonplace.

Another reason for our bewilderment is that digital technology has some distinctive properties — the posh term for them is ‘affordances’ — that make it very different from the GPTs of the past. Among these affordances are:

  • Zero (or near-zero) marginal costs;
  • Very powerful network effects;
  • The dominance of Power Law statistical distributions (which tend towards winner-takes-all outcomes);
  • Technological lock-in (where a proprietary technology becomes the de-facto technical standard for an entire industry);
  • Intrinsic facilitation of exceedingly fine-grained surveillance; low entry thresholds (which facilitates what some scholars call “permissionless innovation”);
  • A development process characterised by ‘combinatorial innovation’ which can lead to sudden and unexpected new capabilities, and an acceleration in the pace of change and development;
  • And the fact that the ‘material’ that is processed by the technology is information — which is, among other things, the lifeblood of social and cultural life, not to mention of democracy itself.

These affordances make digital technology very different from the GPTs of the past. They’re what led me once, when seeking a pithy summary of the Internet for a lay audience, to describe it as “a global machine for springing surprises”. Many of these surprises have been relatively pleasant — for example the World Wide Web; VoIP (internet telephony); powerful search engines; Wikipedia; social networking services; digital maps. Others have been controversial — for example the file-sharing technologies that overwhelmed the music industry; or the belated discovery (courtesy of Edward Snowden) of the pervasive surveillance enabled by the technology and exploited by governments and corporations. And some surprises — particularly the capabilities for cybercrime, espionage, IP and identity theft, malware, blackmail, harassment, and information warfare — have been worrying and, in some cases, terrifying.

But maybe another reason why we are taken aback by the rise of the Internet is because we have been so dazzled by the technology that we have been infected by the technological determinism that is the prevailing ideology in the reality distortion field known as Silicon Valley. The folks there really do believe that technology drives history, which is why their totemic figures like Marc Andreessen — the guy who co-authored Mosaic, the first proper web browser, and now heads a leading venture capital firm — can utter infantile mantras like “software is eating the world” and not get laughed off the stage.

But technology is only one of the forces that drives history because it doesn’t exist — or come into being — in a vacuum. It exists in a social, cultural, political, economic and ideological context, and it is the resultant of these multifarious forces that determines the direction of travel. So in trying to understand the evolution of the Internet, we need to take these other forces into account.

As far as the Internet is concerned, for example, the things to remember are that, first of all, it was a child of the Cold War; that in its early manifestations it was influenced by a social ethos which had distinct counter-cultural overtones; and that it was only relatively late in its development that it was taken over by the corporate interests and intelligence concerns which now dominate it.

Oh — and I almost forgot — there is that enormous elephant in the room, namely that it was almost entirely an American creation, which perhaps explains why all the world’s major Internet companies — outside of China — are US corporations and thus powerful projectors of American ‘soft power’, a fact which — coincidentally — might help to explain current European fears about these companies.

Just for the avoidance of doubt, though, this is not a rant about American dominance. My personal opinion is that US stewardship of the Internet was largely benign for much of the network’s early history. But such stewardship was only acceptable for as long as the Internet was essentially confined to Western industrialised nations. Once the network became truly global, US dominance was always likely to be challenged. And so it has proved.

Another problem with focussing only on the evolution of the network only in terms of technology is that it leads, inevitably, to a Whig Interpretation of its history — that is to say, a record of inexorable progress. And yet anyone who has ever been involved in such things knows that it’s never like that.

With hindsight, for example, we see packet-switching — the fundamental technology of the network — as an obvious and necessary concept. But, as Janet Abbatte has pointed out in her illuminating history, it wasn’t like that at all. In 1960 packet-switching was an experimental, even controversial, idea; it was very difficult to implement initially and some communications experts (mostly working for AT&T) argued that it would never work at all. With the 20/20 vision of hindsight, these sceptics look foolish. But that’s always the problem with hindsight. At the time, the scepticism of these engineers was so vehement that it led Paul Baran to withdraw his proposal to build an experimental prototype of a packet-switched network, thereby delaying the start of the project by the best part of a decade.

Focussing exclusively on the technology creates other blind spots too. For example, it renders us insensitive to the extent to which the Internet — like all major technologies — was socially constructed. This is how, for example, surveillance became “the business model of the Internet” — as the security expert Bruce Schneier once put it. In this case the root cause was the interaction between a key affordance of the technology — the power of network effects — and Internet users’ pathological reluctance to pay for online services. Since the way to succeed commercially was to “get big fast” and since the quickest way to do that was to offer ‘free’ services, the business model that emerged was one in which users’ personal data and their data-trails were harvested and auctioned to advertisers and ad-brokers.

Thus was born a completely new kind of industrial activity — dubbed “surveillance capitalism” by the Harvard scholar Shosana Zuboff — in which extractive corporations like Google and Facebook mine user data which can then be ‘refined’ (i.e. analysed) and sold to others for targeted advertising and other purposes. Although this kind of spying is technologically easy to implement, it could not have become the basis of huge industrial empires without user consent, or without legal arrangements which discourage assignation of ownership of distributed personal data.

One of the most noticeable things about our public discourse on the Internet is how a-historical it is. This is partly a reflection of the way the tech media work — most journalists who cover the industry are essentially perpetually engaged in “the sociology of the last five minutes,” chasing what Michael Lewis memorably described as The New New Thing. As a result, the underlying seismic shifts caused by the technology seem to go largely unnoticed or misunderstood by the public. Yet when we look back at the story so far, we can spot significant discontinuities.

One such, for example, was the appearance of Craigslist in 1996. It was a website providing free, localised classified advertising which started first in San Francisco and gradually spread to cover cities in 70 countries. For a surprisingly long time, the newspaper industry remained blissfully unaware of its significance. But if journalists had understood their industry better they would have seen the threat clearly.

For newspapers are value chains which link an expensive and loss-making activity called journalism with a profitable activity called classified advertising. But one of the affordances of the Internet is that it dissolves value chains, picking off the profitable bits that it can do better than conventional operations. And classified advertising turned out to be one of the things that the internet could do very well: instead of having to wade through acres of small print looking for that used car of your dreams, you simply typed your requirements into a search engine and Bingo! — there were the results. The end result was that newspapers were left holding only the unprofitable, loss-making, part of their value chains.

“The peace of God,” says the Bible, “passeth all understanding”. So too do the valuations of Internet companies. We saw that in the first Internet boom of 1995-2000 — that extraordinary outbreak of what the economist Robert Schiller dubbed “Irrational Exuberance” and which was later christened the “dot-com bubble”. What fuelled the mania was speculative interest in the stock-market valuation of the multitude of Web-based companies (‘dot-coms’) which materialised following Netscape’s IPO in 1995 and which was amplified by the fantasies of fund managers, stock analysts, journalists and pundits. As one sceptical observer put it, what really happened is that “Wall Street moved West”.

The core business model of these fledgling companies was the idea of harnessing the network effects implicit in the rapid growth of consumer interest in the Internet to obtain a dominant market share in a range of sectors. At the height of the frenzy, dot-com companies with few customers, few (sometimes no) revenues and handfuls of employees briefly enjoyed stock-market valuations greater than those of huge companies like General Motors.

The boom followed the traditional pattern of speculative manias through the centuries, and eventually, in March 2000, it burst. In just over a month the total market capitalisation of companies on the NASDAQ exchange fell from $6.71 trillion to $5.78 trillion. In other words, nearly a trillion dollars in value had been obliterated. And less than half of the dot-coms founded in the boom survived the crash.

But here’s the strange thing: the bubble created much of the technological infrastructure necessary to hasten the maturing of the network. When the mania began, some canny observers quoted the old maxim of the Californian gold rush of the 1850s – that the people who made most money in California were not the miners and prospectors, but the merchants who sold them pickaxes and shovels. The modern embodiments of those merchants were the telecommunications companies which in the 1990s invested heavily in building large fibre-optic cable networks and server farms to service the ‘new’ economy that was apparently coming into being. When the bubble burst, these companies were left with apparently unwanted assets, and some went bankrupt. But the infrastructure that they had built remained, and turned out to be critical for enabling what came next.

The interesting thing is that — to those who know their economic history — this is an old story. Brad DeLong points out, for example, that the ‘railway mania’ of the 19th century lost investors a lot of money, but the extensiveness of the railway network that was the product of the frenzy enabled completely new industries to be built. It was the completion of the railway network, for example, that enabled the rise of the mail-order industry — which for two generations was a licence to print money in the United States.

Similarly with the Internet. While the bubble caused a financial crash, it also resulted in a massive expansion in the communications infrastructure needed to turn the network into a ubiquitous public utility — a General Purpose Technology — much as happened with railway networks in the late 19th century. So now the internet is mature and extensive enough to serve as a foundation on which new kinds of innovation – much of it in areas apparently unrelated to information goods – can be built. In that context, it’s conceivable that enterprises like the cab-hailing application Uber, or the room-hiring service Airbnb may turn out to be the contemporary equivalent of the mail-order services of the 19th century: unthinkable before the technology and unremarkable afterwards.

We’ve taken a long time to get here, but we’ve made it. Now all we have to do is figure out how to deal with it. Which is why I say that the arrival, not the journey, matters.

Thank you.

Resources, resources, resources

Here’s an instructive story:

Just three months after launching Facebook Live to all users, Mark Zuckerberg decided to go big, realizing in a February meeting that the company should make Live a top priority. A BuzzFeed story on Live tells us what happens next, quoting Facebook Media’s product lead Fidjij Simo:

“The original Live team was composed of only a dozen or so people. But the vision laid out for the product at that February meeting would require more than 100 engineers to build. ‘The meeting was on a Thursday, and on Monday, [Facebook Media engineering lead Maher] Saba and I were standing in front of 150 engineers,’ said Simo.”

From 12 engineers to 150 in less than a week. That’s the new pace of the media business.

Now here’s another instructive question: how many engineers are working on the iPhone camera?

50?

150?

Nope. At the moment, there are 800 engineers working just on the camera.

Which leads me to wonder how many people work in the R&D divisions of Nikon and Canon?

See what conventional companies are up against?

flickr_stats

At present, the most popular cameras among Flickr users are all iPhone models. By a mile.

Improbable? I don’t think so

There’s an interesting new company in London with the improbable name of, er, Improbable. It’s funded by (inter alia) Andreessen Horowitz and it’s built a distributed operating system (called SpatialOS) that enables the operation of huge systems to be realistically simulated.

This has obvious applications in areas like gaming and VR, but the video shows a really intriguing example which has little to do with either. According to the company’s blog,

a team of two came in from the British government1 to explore our technology. Their goal was to build a realistic simulation of the internet so that they could take a look at its “structure”, or in other words, the vast number of connections between computers and networks that make up the World Wide Web. With the internet under attack from a variety of sources, it’s critical they can see its weak spots, to figure out how to protect it.

According to the blog post, Improbable engineers and the visiting spooks were able to use SpatialOS to build a simulation model of the entire Internet backbone in just three days.

“Not only did we demonstrate a dynamic model of BGP routing at scale, we also produced an interactive visualisation where both ASs and the connections between them can be created or destroyed, observing dynamic routing, cascade failures and new route propagation across the network.”

This could be really useful, because computer simulation is one of the few tools we have for trying to understand the behaviour of very large complex systems. But most of the simulation tools we currently have run out of capacity when the systems are as large and complex as the Net. We needed something more powerful. Maybe SpatialOS is it.

They’re looking for ‘developer partners’ btw.


  1. Presumably from GCHQ or the Cabinet Office.