How fast is the world changing – and how fast are we adapting to it?
Executive Summary: Faster than we think, and not fast enough. At least those are my takeaways from The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, by Erik Brynjolfsson and Andrew McAfee. an interesting book that’s simultaneously uplifting, optimistic, terrifying and worrying. And definitely worth a read.
I’m not sure that’s what Messrs Brynjolfsson and McAfee intended as the main takeways – and there’s certainly plenty else in there about the economics (and inequalities) of the machine age, the spread of robotics, data and algorithms into our lives, the likely policy implications, and all that sort of thing – but at least that’s how I read it.
Not that it’s necessarily bad news – in fact, one of the best frames they put on the new digital age that we’re in is around the notion that, just as the industrial age freed us from the constraints of animal and human muscle power, this revolution is set to augment the human brain in ways we can’t yet really understand.
As they point out, a whole slew of new capabilities have emerged in the last few years that were effectively dismissed as essentially impossible in the near term less than a decade ago – self-driving cars, effective voice recognition, human-like text generation, visual recognition systems, and so on. Even machine translation services, which are in many ways laughably bad, are a minor miracle in that they even exist.
What’s driving all this? Increasingly powerful computing, decreasingly expensive sensors and processors that are proliferating in devices and objects, and the masses of data they’re generating every minute. All that data isn’t necessarily useful – and certainly will be misused at some point – but it opens up lots of potential new capabilities. But first we have to imagine them.
Look at Waze, for example, which took the ubiquitous location – and hence speed – data available on phones to build real-time traffic maps. Before that, we had various mapping apps, that while useful, basically took what was static information and put them on mobile devices and added search and other capabilities. Very useful, but much more an extension of existing systems. It took the folks at Waze to reimagine what they could do with completely unrelated data. So what other tricks are we missing? (What if we used cellphone location data to help the blind navigate around other people in crowds?)
And, of course, I’m still plugging structured journalism.
Frederic Filloux, in a recent Monday Note, pointed to a host of potential news applications that could be built off the data in our mobile phones, going beyond simply the notions of shorter stories/small screen-friendly articles/quick updates that dominates a lot of the discussion of mobile apps.
Consider Google Now, the search engine’s intelligent personal assistant: It knows when you are at work or at home and, at the appropriate moment, it will estimate your transit time and suggest an itinerary based on your commute patterns; or take Google Location History, a spectacular — and a bit creepy — service for smartphones (also tablets and laptops) that visualizes your whereabouts. Both Google services generate datasets that can be used to tailor your news consumption. Not only does your phone detect when you are on the move, but it can anticipate your motions.
Based on these data sets, it becomes possible to predict your most probable level of attention at certain moments of the day and to take in account network conditions. Therefore, a predictive algorithm can decide what type of news format you’ll be up for at 7:30am when you’re commuting (quickly jumping from one cell tower to another with erratic bandwidth) and switch for faster reads than at 8:00pm, when you’re supposed to be home, or staying in a quiet place equipped with a decent wifi, and receptive to richer formats.
By anticipating your moves, your phone can quickly download heavy media such as video while networks conditions are fine and saving meager bandwidth for essential updates. In addition, the accelerometer and internal gyroscope can tell a lot about reading conditions: standing-up in a crowded subway or waiting for your meeting to start.
What all this takes is less a technology background and much more a willingness to rethink what we do. Certainly this revolution – unlike the industrial age – allows for much faster recombination of capabilities and cheaper experimentation, since much of it is based on moving digits, and not atoms, around. But how well-equipped are we to do this, rather than simply use the new capabilities to do what we currently do more efficiently. Much of what’s been done so far, note Messrs Brynjolfsson and McAfee, is essentially “paving the cow paths…”
…as opposed to rethinking how the business can be redesigned to take advantage of new technologies.
The analogy with the introduction of electricity a century ago is one that’s been much written about (including here). Basically, it’s about how electricity was first seen as a cheaper, more stable substitute for steam or water power, which it certainly was. But it was only after factory owners understood that it could power small machines and allow for the complete reorganization of the factory floor were the real productivity benefits – and other revolutionary creations – of electricity were unleashed. The scary thing is how long it took – and more scarily, what it took:
Only after thirty years – long enough for the original managers to retire and be replaced by a new generation – did factory layouts change.
In other words, the older generation simply couldn’t imagine a different factory floor. I don’t think we can wait that long. For one thing, we’ll cede the future to others, who may not share some of the public interest ethos that should lie at the heart of journalism. For another, I don’t think we have enough in our 401(k)s.