Great minds like a think. And Demis was just getting started. Imagine, if you will, a line of people consisting of a cross section of human beings going from the most selfish, immoral, dishonest, sexist, racist, narcissistic, mindless, petty and cruel at one end (I know, you’re way ahead of me) and the absolute polar opposite at the other, a human being who embodies all of our species’ best characteristics; altruism, empathy, kindness, honesty, and who is frighteningly intelligent with mankind’s best interests at heart. If we looked at that very opposite end from the reigning champion of sheer awfulness (unnamed for legal reasons), we’d find Demi Hassabis. While competing at the very highest levels of chess at the age of seven in a hall containing scores of contestants, each with an equivalent amount of brain power as he possessed, little Demis had a mighty thought. What if he could assemble the finest minds of any living generation in order to create something extraordinary and not waste those giant minds’ time playing a board game…

And what tool would Demis harness to create such extraordinary breakthroughs? Is there even any doubt? Artificial intelligence, we keep being told, is a game changer. Cinema and literature has prepared us for the rise of the machine, a Terminator-esque march of armed, skeletal robots crushing skulls underfoot. As no one knows what’s coming or what AI might mutate into, there are legitimate fears of our machines becoming our unsympathetic overlords unless there are rigorous, global safety protocols in place, legislation that everyone agrees to adhere to. Yeah, like that’s ever going to happen. The safety protocols will be in place. I’m just not sure the human animal will take any notice of them. Like the internet, offering useful information as a throw away side line to providing pornography to the masses (its principal raison d’etre which doesn’t say a great deal about mankind), AI’s modest side line will be solving big problems for the benefit of the planet and its inhabitants. Its main use, in a depressingly far greater percentage of cases, will be to make the very few, very rich or richer. But that tiny percentage who have mankind’s and the planet’s best interests at heart will be showing us what AI could mean in our (or its) next evolutionary step. AGI (‘G’ for ‘general’) is being created in exponential stages to solve some of our most insoluble problems. At the forefront of this mankind-friendly project is a man whose unassuming features contain a mind that is so above and beyond accepted norms, it must have seemed like the most daunting task to make a film about him. How do you do justice to a genius and the über-gifted team he assembled? How do you show the potential of such mind power when the scaffolding and architecture of their achievements is beyond most people’s comprehension? Kudos to director Greg Kohs for attempting it. But even a genius needs patronage to fully blossom. Imagine Michelangelo’s Sistine Chapel ceiling without the might of the Catholic Church’s riches behind him. The celebrated Sistene Chapel’s door frame? Demis and his team were taken on by Google Deep Mind with Demis setting the rules for what would follow given the success or failure of the research. I would normally be wary of Google, as I would the Catholic Church, but even if the company is at the infamous top spot for data harvesting without individuals’ permissions, it has massive power and facilities that Demis and his team could exploit in their research.

The one entry level ‘in’ the filmmakers had which everyone could understand was that the development of the AGI began with encouraging the code to learn how to play and win video games starting with, of all things, Pong. It doesn’t get more basic than that. With the goal of increasing its own side’s score, the AI suffers heavy losses until it figures out exactly what it has to do to win and then replicate that behaviour with unerring accuracy. From baby steps, the AI – always a terrific student – devised winning strategies for all manner of computer games even showing tactics to win even the game designers hadn’t thought of. The granddaddy of board games, the 2,500 year old Go, seems simple enough. On a grid of 19 x 19 squares, black and white flat stones are placed on the squares’ intersections. The aim is to ringfence your opponent’s territory. It sounds fairly straightforward but the number of legal board positions in the game has been calculated to be approximately 2.1×10¹⁷⁰, which is far greater than the number of atoms in the observable universe, which is estimated to be on the order of 10⁸⁰. “Space,” as Douglas Adams so wisely informed us, “…is big. You just won't believe how vastly, hugely, mind-bogglingly big it is. I mean, you may think it's a long way down the road to the chemist's, but that's just peanuts to space.” So… the near infinite complexity of Go makes it a perfect learning tool for AGI.

Greg Kohs first covered this aspect of Google Deep Mind’s march into the future in his film AlphaGo, named after the AGI designed to play Go better than any human being. It competed with the then Go World Champion, Lee Se-dol, a South Korean player with a well-earned reputation for cognitive brilliance. No spoilers in this review. You could almost regard AlphaGo as a feature-length extra on a potential The Thinking Game Blu-ray. The team’s involvement with AlphaGo led to the their work in molecular biology and an attempt to solve the knotty problem of how a string of amino acids fold themselves into the perfect structure for the job they evolved to do. In 1994 there was even an organisation set up, Critical Assessment of Structure Prediction (CASP), which started a biennial competition to see if anyone could predict, with a high percentage of accuracy, exactly what shape these proteins would fold into. It’s like knowing every component part of a Ferrari down to the smallest nut and creating code to figure out the prospective shape of the sports car. Of course it is nothing like that but more complex analogies escape me.

Shot and edited beautifully by director and DOP Greg Kohs (with a team of camera operators) and Steven Sander respectively, the film looks terrific and whizzes along at the perfect pace. You’re never left in the dust and never stuck in one place too long to labour a point. It’s quite the balancing act. What The Thinking Game does with such great effect is make this monumental scientific adventure into a very human story as all stories ultimately are. You don’t have to understand the code that resulted in AlphaGo’s infamous ‘move 37’ in a tense Go match but seeing how the team responds to it is all you need to know to be thoroughly entertained and inspired by what the primate known as the ‘human being’ really can achieve. It is especially inspiring given the current state of the world turned that way by those unfit to run it. Demis and his team reveal an aspect of our higher ambition that doesn’t involve profit or recognition, just helping people and making the world a better place. Before Elon Musk mutated into an unrecognisable, conspiracy-pushing, power-obsessed deceiver, he was also a hero to many. Demis spoke to him, before his metamorphosis as a possible sponsor. He rounded off a presentation with one of Musk’s own ambitions… “His mission is to die on Mars… but not on impact!” Who said geniuses have no sense of humour?

You want some light in the darkness? See The Thinking Game.

* https://cbmm.mit.edu/about/people/hassabis