Peter Kawalek: 20 years in tech
I was a PhD student in Manchester when a robotics lab was opened in the Department of Computer Science. It was 1997, or thereabouts. The robot itself was not as impressive as I had anticipated. It looked like an upturned waste-paper basket with wheels. Its task was to learn to navigate through the department’s first floor by itself. It had no sensory discrimination over a transient object like a human and something permanent like a wall. This meant that we humans were responsible for stepping out of its way should it approach. Quite regularly one would see it, wheeling its little way through the corridors, trying to learn the layout of the corridors. I do not recall if it was based upon a rule-based AI or some more extensive Machine Learning algorithm. Either way, the little robot was frequently found stuck in a doorway or at corridor’s end. One could helpfully nudge it on its way as one returned from coffee.
In 2004, Levy and Murnane published ‘The New Division of Labour’. In it they speculated that driving was one of those sentient tasks that would always be beyond the wit of machine intelligence. No computer, they proposed, would ever navigate a truck through a busy intersection.
In 2010 Google launched its autonomous car project. Waymo, the company founded by Google to bring autonomous cars to market, was launched in 2016. This year, it was announced that autonomous trucks will be tested on the UK’s roads. Tesla already has a semi-autonomous mode of hands-free driving for its cars and has over 1.3 billion miles of experience. The machines learn all the time that a Tesla moves.
Tech is breaching the dam of expectations. It is certainly ahead of my expectations of 1997. I did not expect trucks to navigate motorways within 20 years of that little robot, and it is ahead of Levy and Murnane.
The automotive industry is especially important as an indicator of the advance of ‘tech’ because we understand the complexity of taking a car through Houston, as Waymo does. We also understand the social significance of a day when the disabled and the old are able to take autonomous vehicles, when children take autonomous vehicles, when the rest of us convert to autonomy, and when we stop buying cars and prefer pay-as-you-go.
We intuit that vehicles that are more road-efficient, more environmentally friendly, and more socially inclusive, are hugely important for society. They might also mean the end of some of the great car companies of today. Most likely, the oil-economy will take another very serious hit. That will change geopolitics.
Could this prospect be bigger?
The term ‘accelerationist’ is applied to the thesis that innovation is speeding up. The accelerationist argument observes that there are combinatorial effects between innovations (the autonomous car depends on Machine Learning, Big Data, the Cloud, Internet of Things, faster chip speeds, satellite navigation and batteries).
These combinatorial effects tend to have the effect of increasing the overall rate of change as they multiply the possible number of applications of any given piece of kit. Though mainly associated with futurists and tech-investors, it seems to me that in many circumstances this accelerationist argument is sensible and realistic. There is evidence. As we have progresses from an upturned waste basket with wheels in 1997 to Tesla and Waymo in 2017, what is in store for the next twenty?
Once, and laughably in retrospect, we thought of ‘tech’ as a sector different to other ‘traditional’ sectors like automotive, or energy, banking, law, or, indeed, education. Now we know tech as more or less everything outside of the artisan bakery. Find me an oil executive, a banker, lawyer or manufacturer who is not talking about ‘tech’ and I will find you someone who is close to retirement.
Since the little robot of 1997, we have seen the rise of Amazon, Facebook, and Google. Steve Jobs returned to Apple. Microsoft won the browser wars but lost on mobile. Social media, cameras, telephony and tracking devices began to crowd our lives. Teachers began to outsource homework to websites. The BBC, newspapers, record companies, malls, advertisers and banks all lost key positions. Cyber-crime was born.
Think about the systemic effects of the cookie – that little bit of spying functionality that we accept on our computers. It is not an exaggeration to say that the cookie begat new forms of commerce and in some ways changed the world. More and more we are beginning to package and price attention itself as Google and Facebook run billions of auctions for personalised advertising.
Probably we should be reaching for Schumpeter, Kondratieff, Kuhn…
Yet, with fairness, one cannot say that British business schools have been at the cutting edge of this change. Scholars have established some good reflective positions, some valuable insights, but we are not associated with anything like the river of ideas and connections as Stanford or Sloan. There are obvious reasons like geography and funding: this is primarily an American revolution. There are also other reasons that are less obvious but still very important.
I have been reflecting on how William Barton Rogers sowed a commitment to ‘learning by doing’ into the primary constitution of MIT. Here in the UK we actually have a great tradition of engaged and action-based research, indeed possibly we have the world’s best academic works on this kind of practical theorising and value creation.
So, in short, we should be thinking about the methodological challenge of researching the fast-changing, turbulent, multifaceted , tricksy phenomena of digital. Herein is an opportunity and portal for Loughborough.