Back in the 90s I was fascinated by IBM’s Deep Blue project – which aimed to create a computer program worthy of defeating the world’s greatest Chess champions – Garry Kasparov’s loss in particular was a huge moment in what was seen then as artificial intelligence, but with today’s perspective just some brute-force programming.
AlphaGo (now acquired by Google) is today’s equivalent, with two key differences. This time the game is Go – a 2500 year old Chinese game with far more complexity than chess. Secondly, AlphaGo was designed with deep learning and improves with repeated plays.
The doco makes for great viewing (especially being stuck inside right now) and it’s free to watch on YouTube! It’s fascinating to watch Lee Sedol battle with the emotions of playing against a machine and everyone LOSE THEIR MINDS with move 37.
We’re not just teaching computers any more – they’re teaching us. And what’s more – it’s coming for my vfx industry with Netflix’s Archer being one such example (no, not that one). I’m not one to proclaim changes like this to be doom and gloom for the artist. In my experience, clever tools only free up artists from mundane activities and leave them to focus on the art. I always say “if something can be automated, it should be”, and advances in tech and tools have only led to a greater volume of work. That said, it will be an interesting shift.
