Latest Blog Post

I wrote an introduction to the algorithm in the previous blog here.

But I wanted to try to convince you of how artificial intelligence and machine learning is going to change the way you live, the way you work, whether you like it or not.

So, taking references from Harari’s book 21 lessons, this is the story that made me sit up and think even more about this subject.

For a long time, chess was considered to be one the pinnacles of human intellectual achievement. The strategy and intuition and creativity required in this game was beyond that of machines or anyone else apart from us.

Everything changed in 1997, when deep blue (the IBM supercomputer at the time) beat Garry Kasparov, for the first time.

Everything changed from then.

Since 1974 there has been a world computer chess championship, but after 1997 all of the computers got really good, with hundreds of years of human compute of human chess activity embedded and software written by humans to play chess, who could play more efficiently and more effectively, if would take another 19 years though, from machine learning and artificial intelligence to enter the world of computer chess, in 2016 Google created Alpha Zero.

Alpha Zero was not a chess computer, it was a machine learning device, which was able to teach itself how to play chess.

Google challenged, the world computer challenge champion of it’s time, for a hundred matches against Alpha Zero.

Alpha Zero won, 28 and 72 were drawn. The world computer chess champion won them.

This may not seem very impressive, but it does begin to seem more impressive and even sinister when you learn that Alpha Zero taught itself to play chess in 4 hours.

Write the code to learn, press the button and watch it learn.

Once machine learning is properly introduced the machines will learn quickly, more quickly than we can even fathom.

So, where does that leave us? How does that impact on us? Because I don’t play a lot of chess.

You have heard or have read recently, about Babylon health - no waiting rooms, no waiting, this is the start of the AI Doctor.

You will sit back and rail against that and suggest to me that machines can never replace doctors, but consider this.

The AI Doctor is on the network, networked to however many other AI Doctors there are immediately.

It learns.

Babylon health has been criticised for the mistakes that it has made and the problems that it’s encountered, but the network will learn.

It will learn a lot faster than a network of doctors and while it may not be perfect very quickly, it will be much better for the lowest performing doctors.

Sooner it will be better than the mid-ranged guys, and soon after that, well…

This has been tested in London, where you can’t get a GP, and this is the model for testing machine learning and artificial intelligence.

First, put it somewhere, where it isn’t that thing, then show how good it is, so that the people who do have access to the human version, then want the machine version.

Imagine the AI doctor in Tanzania, where there are no doctors.

This is not supposed to be frightening, not supposed to be a doomsday scenario, imagine how that will impact on people like Matt, who is working in my practice, this week and just about to enter final year dental school.

In 10 years’ time the way Matt does dentistry with augmented reality and machine learning associated, will be unrecognisable, to the way that we do it today.

Blog Post Number - 2072