The weekly download: No 43
Real world thinking about AI's future: nurturing the future inside the Xiaomi ecosystem; why owning a self-driving car doesn't make sense; and lessons from chess about human/ machine hybrids...
I am unusual in the tech world because I did not grow up loving sci-fi movies and video games. So my view of the future is based on what I have learned working with real businesses, not drawn from stories in books and on screens.
My first article this week is Grace Shao ’s post about Dreame is a good example. Dreame is a Chinese company that started supplying robot vacuums to Xiaomi and now has an ambition to reinvent pretty much every category of consumer electronics.
The story is part entrepreneurial hustle and part customer led vision. The entrepreneurial part starts inside the ecosystem of another fast growing business, Xiaomi. How many other supply chains in China are nurturing future ambitions this way?
The main business insight for me is the way the founder, Yu Hao, sees AI as an opportunity to reinvent hardware. Remarkable vision coming from a background in such mundane devices, made more powerful because that history means Dreame really understands what users want.
Specifically, there is none of the popular delusion about humanoid robots here. The way forward is AI and robots designed to do jobs. Not designed to look like the people who currently do these jobs.
Another hint of a more realistic future in a post from Understanding AI which explains Why it might not make sense for you to own a self-driving car. It explains the economics behind a high-tech vehicle that requires regular maintenance and updates to both hardware and software.
Economics are one thing. Emotion will decide whether self-driving cars take over our roads or remain an idea that never catches fire.
Driving creates an emotional connection to the vehicle. Once you break that, how many people will care about ownership? If every car is self-driving, will humans even be allowed to drive on public roads? Could driving a car become a hobby restricted to a few racetracks?
Paul Willmott writes about a closer future in The Centaur Age Is Here. Which of Yours Are Lame? His chess analogy is well chosen. His take on the immediate future of AI as a hybrid of human and machine is credible.
I would extend the chess analogy. El Pais recently published an interview with Javokhir Sindarov, winner of the world championship candidates tournament earlier this year. In it he explains why computers are not only better at chess than humans, they have changed the way grandmasters play the game:
“I’ve never looked up games of great champions of the past on my own”...”I can summarize it like this: I’m always calculating, and when I have to make a decision without time to calculate everything I’d like, I rely on my intuition. Very concrete calculation is the basic element of modern chess.”
Willmott’s article talks in some detail about the way to set up AI/ human combinations for success, followed by 5 specific questions Boards can ask to steer AI development.
I am all in favour of questions but steering is the wrong way to frame the Board role. And the detailed prescriptions earlier in the piece ignore the fluid reality of AI today. We don’t know what this will look like and the best approaches will change radically and fast.
The best course for Boards and management is to encourage curiosity, watch what works and lead through evolution.