Business Recoded … Whilst we marvel at extreme feats of human performance, we also know that technology has the potential to outperform humanity.
March 19, 2020
In an exclusive extract from my forthcoming book Business Recoded, meet one of the most inspiring concepts shaking up today’s world. DeepMind embraces the opportunities of relentless change, the incredible latent power of artificial intelligence, and the possibilities of technology to create a different, also better, future for humanity. In the book, I explore the stories of many of the world’s most fascinating businesses and innovations right now, and develop 49 codes that help you redefine the future of your business, and yourself.
The Future Code of DeepMind
The ability to process huge amounts of data at incredible speeds, to learn through repetitive process, and to harness the strength and agility of robotics challenge many of the ways in which humans used to excel.
The game of chess has long served as a benchmark for artificial intelligence researchers. John McCarthy, who coined the term “artificial intelligence” in the early 1950s, once referred to chess as “the Drosophilia of AI” comparing it to the way in which the fruit fly is used to understand genetics.
In 1996, IBM’s Deep Blue supercomputer embarked upon a series of chess games against Garry Kasparov, the world champion. Deep Blue eventually beat Kasparov, marking the first time a machine had defeated a world champion.
Within a few years computing technology was consistently beating chess grandmasters.
However, AI developers knew that they needed greater challenges, searching for more complex games to test their increasingly sophisticated algorithms. They turned their attention to the ancient Chinese strategy game of Go, which is both deceptively simple to play, yet extraordinarily complex to master.
The game was invented in China more than 2,500 years ago and is believed to be the oldest board game continuously played to the present day. It was considered one of the four essential arts of the cultured aristocratic Chinese. Go has a larger board than chess, a 19×19 grid of lines containing 361 points, and therefore with many more alternatives to consider per move.
It took another decade of machine learning development until scientists were able to create a truly competitive AI-based Go player.
In 2014, a team at London-based Deepmind Technologies started working on a deep learning neural network called AlphaGo. Two years later a mysterious online Go player named “Master” appeared on the popular Asian game platform Tygem. The mysterious player dominated games against many world champions.
Eventually it was confirmed that the “master” was in fact created by DeepMind, since acquired by Google, and now a subsidiary of Alphabet.
The master was replaced by a grandmaster in 2017. AlphaZero, an enhanced version of the original system, embraced an even more sophisticated algorithm designed to learn as it progressed through games. The system simply plays against itself, over and over, and learns how to master whatever game it has been programmed to work with. Searching through 80,000 positions, a fraction of what other predictive software had used, it had perfected the game in 24 hours using a AI-type of intuition.
AlphaZero achieved two things: autonomy from humans, and superhuman ability. Scientist and futurist James Lovelock calls this “the novacene”, translated as “the new new” in Latin and Greek, where a new form of intelligent life emerges from a human-initiated AI-based machine into one which no longer requires human intervention.
He calls AlphaZero, and other such beings, cyborgs.
In his book Novacene: The Coming Age of Hyperintelligence, Lovelock suggests that AI-based entities can think and act 10,000 times faster than humans (and to put that in perspective, that humans can think and act 10,000 times faster than plants). He then reflects that maybe AI-based life would be rather boring, considering that a flight to Australia using physical transport would currently take 3000 AI-based years.
The real point of a cyborg, a term first coined by Austria’s Manfred Clynes to describe an organism as self-sufficient as a human but made of engineered materials, is that it is able to improve and replicate itself.
Which quickly takes us to a future beyond what Hungarian John Van Neumann called “the singularity”, the point at which technological growth becomes uncontrollable and irreversible. Both physicist Stephen Hawking and entrepreneur Elon Musk have warned of the profound implication of autonomous AI.
Of course, we already have many devices which learn and improve continually. Take Google Maps, for example, which constantly learns from all its users about realtime traffic situations, and the more users it has the better the information becomes. Or consider Google Nest, an intelligent thermostat which takes control of the temperature in our homes. For now, they are useful tools, to help us live better.
In the meantime, DeepMind continues to explore the future. StarCraft, considered to be one of the most challenging Real-Time Strategy (RTS) games and one of the longest-played esports of all time, has emerged by consensus as the latest “grand challenge” for AI researchers around the world.
© Peter Fisk 2020. Business Recoded will be published in September 2020.