Alexa is Good, Almost
On my drive home today I attempted to ask Alexa, via Amazon Music, to play California Love by Tupac. It was a beautiful autumn day… I rolled my windows down… and just needed something to “bump” with. Alexa, who generally does okay, met me halfway. I only knew California and Tupac. Just by that alone she understood I wanted California Love by Tupac. And yet, she still failed by delivering me a (remix) version instead. I’m not going to get into it here, but rarely am I interested in remixes or live versions… Someday Alexa will understand me.
AI is Going to be Big
I went to a banking conference in New York back in September, called Finovate Fall. New York was awesome, the conference was good, etc… However, a full day was spent on artificial intelligence, where we are and where we will be in 5-10 years.
Probably my favorite presentation of the whole conference was by Tim Urban when he talked about the road to super intelligence. It really peaked my interest and I have been thinking about it off and on ever since.
The cliff notes version is that today we have what is called “Narrow Artificial Intelligence,” this helps us select new movies and tv shows to watch in Netflix. “You liked that? Well, other people, just like you, liked this!” This kind of AI is all over the place right now.
Next you have “General Artificial Intelligence.” We’re not here yet, but we are getting closer. What does this look like? Well, right now websites can tell that we’re not a robot by asking us to click all the pictures that have a dog in them. Sometimes they throw in a wrench where a cat looks similar to a dog, but only a real person could tell. If a computer was “GAI” then it could tell the difference too.
It’s funny how the things that humans are good at, computers suck at and vice versa. There is a lot of research and money going into this area of AI. Which leads to “Artificial Super Intelligence!” This is where a computer can not only think and act like a human, but also reprogram itself to work harder, better, faster, stronger… just like Daft Punk.
5 years ago I thought that actual artificial intelligence was impossible. But in just another 5 years it will be everyday life. I didn’t think any computer could actually pass the Turing Test, but once I started thinking about it… lot’s of computer interactions already do, and have for a long time. One example is ELIZA from way back in the 60’s. It simulated a Psychologist by taking whatever the user typed and saying it back to them as a question, more or less… Users who knew it was a computer program still got really into it and even asked for privacy from the researchers to talk to their “chatterbot.”
Star Wars
So, on my drive home, after my Alexa interaction, I got to thinking about what it will be like in 5-10 years to interact with artificial intelligence. Where will they be and how will we talk to them. I think an obvious place with all the “no texting while driving” rules will be in the car.
AI makes it easier to interface with technology because instead of them needing to be coded to transfer information every device is speaking the same language. When we sit down in our car, our phones will automatically sync with the car. When we want a song… we just ask for it. When we get a text message… the computer reads it to us. And, if it was paying attention to the last time we talked to the sender on the phone, maybe it mimics their voice as it reads the text message?
AI is really meant to make our interactions with technology more seamlesss. we could completely be rid of keys. As we walk up to the car, we could just say “hey… Honda Odyssey… open up!” Then it authenticates us as the owner of the vehicle through voice biometrics. When we sit in the seat, we could just tell it where we want to go. Or, if driverless cars still aren’t ready it could just pull up our destination on the GPS.
This is when I started thinking about Star Wars. I mean, it usually isn’t far from my mind… But, the scene where Luke is talking to R2D2 flashed across my minds eye.
Luke, looking thoughtful, suddenly makes a decision. He flips several switches. The stars shift as he takes his fighter into a steep turn. The X-wing banks sharply and flies away in a new direction.
The monitor screen on Luke’s control panel prints out a question from the concerned Artoo.
LUKE (into comlink) There’s nothing wrong, Artoo. I’m just setting a new course.
Artoo beeps once again.
LUKE (into comlink) We’re not going to regroup with the others.
Artoo begins a protest, whistling an unbelieving, “What?!”
Luke reads Artoo’s exclamation on his control panel.
LUKE (into comlink) We’re going to the Dagobah system.
Luke checks his readouts and makes a few adjustments. He rides along with only the soft hum of the instruments to break the silence. Finally, Artoo chirps up.
LUKE (into comlink) Yes, Artoo?
Artoo utters a soft, carefully phrased stream of whistles.
LUKE (into comlink, chuckling) That’s all right. I’d like to keep it on manual control for a while.
THIS IS HOW: I imagine I’ll be talking to my car in a few years…
But, Wait… Aren’t Computers Going to Enslave Us?
After I started thinking about Star Wars my juices really got flowing on what all interactions with AI will be like. For 50+ years Hollywood has prophesied the dangers of allowing computers to become sentient. A Space Odyssey, The Terminator, The Matrix and many many more show that computers are out to get us. But, Star Wars shows AI differently…
The droids in Star Wars are extremely loyal to their owners, you just have to make sure you are the owner. They take pride in being useful and flexing their abilities, they have strong personalities. But, most important they hate being in the foreground, they don’t seek for power, but rather to serve.
I think as we pioneer new bots and AI we should always remember C3PO and R2D2… the pattern for the model AI. Like a loyal dog.
And then I pulled into the garage. That’s all for today. Thanks!
TTFN, tata for now
Mark Freeman
thefreemanblog.com 