I found this article in the Guardian a few days back.
I am quite impressed. I remember when the first chess-computers came out. I am not a chess-player at all. I know what the various pieces do and that’s it. I do know various chess-fanatics and they would keep telling me their utter amazement (and enjoyment) that these chess-computers were just becoming better and better. Later IBM took on various chess-champions, but pretty much always lost, until finally IBM’s Big Blue beat Garry Kasparov in 1996.
There is a pretty decent documentary about it, called “Game Over, Kasparov and the machine”.
Although IBM didn’t play fair, through machine learning, computers have become better and better at figuring complex things out. This is particularly clear in translation programs, text to speech software and voice-assistants. Computers/programs can “learn” human language better and better. This allows us to speak to them, give them commands, ask them stuff, but it also allows them to learn. Most “knowledge” is stored as written word. Scientific papers, university courses, court-documents, medical-research, you name it, it came from academic science books.
In 2011 IBM’s Watson won the first prize on Jeopardy, a quiz-show, competing against human contestants.
Since then Watson’s abilities and potential has only grown.
Google Deepmind’s AlphaGO recently beat the world champion Go player.
Go is an exceptionally complex game.
But the machine beat the champion quite easily.
You should check out Watson and Deepmind on Youtube. Deepmind has its own channel.
This poker victory is really quite something.
Being able to read other player’s behavior/style of play is quite a feat for a computer.
I have mixed feelings about AI. I like the idea of being able to communicate with a computer using my own, real language. I know I already can, but I have no real use for it. I like to type. My daughter, a teenager, does like to speak her searches and commands, she does so on her phone as well as on her Chromebook. I like the idea of telling my computer to play album <x> from artist <y>, but (again) I have no real use for it. I like to click and adjust and fiddle with the volume, while I am doing other things on my Chomebook.
So yes, I like the potential and power behind it, but I am not using any of it (yet).
Actually that’s not true. I am “using” loads of it already: traffic systems, the energy-grid, Google searches etc. A lot of software and systems have already become smarter using machine learning, predictive algorithms and neural networks and such. So I am using it, I am just not consciously and actively engaging with any AI yet. I know I will in the future, as my daughter will use those systems more and more, so will I. (You gotta keep up, old man! eds.)
The speech assistants will be wonderful for people with muscle or bone disorders, anybody who has trouble typing will be thrilled.
I like the idea of automated cars. People drive like idiots and if all traffic was automated, we’d end congestion, parking problems, accidents, pollution… and for people who have long commutes – we’d free up huge amounts of time where they can read, study, work, hang with their kids, all while being driven, instead of being all stressed going 100 km/h in the dark, in shit-weather, with thousands of other idiots going similar speeds, all trying to make it to some remote destination on time.
As a visually impaired person, I’ve never had a driver’s licence.
There is a video online of a blind person being driven in one of those cute automated Google cars.
It brings tears to my eyes.
If they were cheap, or for rent and really looked like this?
I’d love to ride around in one.
Here’s where the problems begin.
Automated cars means automated transport which means a shitload of people (chauffeurs/drivers) would lose their jobs and they’d be ill-equipped to do anything else.
As computers become smarter, more programs will take over “work”.
Capitalists love this idea.
It’s called the Gig-economy.
The Gig-economy is (of course) also a gorgeous piece of corporate propaganda.
Our bills and taxes will stay the same, regular, monthly.
But our income and our children’s income will not.
Steady bills and unsteady income is a recipe for disaster.
A fucking monkey can work out the math, but alas, it’s coming and there’s nothing you can do about it.
Because of computers.
I am doubtful that will happen. The inevitability of it seems rather fatalist.
And whenever capitalists and rich fuckheads really love something and tell me it’s inevitable?
I tend to distrust the entire endeavour.
Another aspect that isn’t considered fully is more philosophical and ethical.
As we create smarter and smarter machines, we will at some point be able to create a superintelligence.
Science is hard at work, in labs across the world, trying to achieve just that.
I encourage you to take a look at Sam Harris, speaking briefly about AI at Ted Talk.
And check out his podcast below.
And honestly, even with all the dangers involved, if one of these came out?
I’d buy one !
So… science-people…? When can I buy a HAL9000?