First generation AI captured the imagination. Expert systems were the direct result of 1st Gen AI. Most of us wouldn't even call that AI. Expert systems are simply the evolution of getting searchable databases and expert procedures committed to a checklist in software. Anyone one who was doing things right, without a computer, had a checklist.
Second generation AI machines can learn, but only thanks to Bayesian statistics and neural networks. But what are they really leaning? Plenty of apps are out there which can give you a feel for how well this type of machine learning works... Take a picture of yourself and have the app search for painted portraits of historical figures. Who do you look like? George Washington? Marie Antoinette? There are thousands of paintings in the archive... But what did we learned in computer science 101? Garbage in, garbage out... Nothing has changed. You train your AI on the wrong data set, and you get garbage. Friends of my would argue this is all too human. You watch Fox News and you train you human brain on garbage. You get a garbage President. Fair enough. At the risk of upsetting my daughter...the problem with the millennial brain, among other things, is that you never know what data set they've been training on. They have an unlimited number of data sets to train on...much of it garbage. Most of us growing up only had our parents to mold our minds. The problem with these unlimited data sets is that some of it is rooted in truth but then it quickly diverges someplace else. If you've had any experience or understanding of how to think critically, maybe you can sort it out...but even the best of us have been fooled by garbage.
A lot of people much smarter than I am have been considering Third Generation AI. What do they have to show for it? Nothing. No one can even predict what Third Gen AI will look like. Some say quantum computing? Nah, that's no real change to the information architecture that will support it...it's all 1's and 0's in the background. It will be the same foundation of logic. No one is going to want to recreate decades of digital logic developed behind the current digital age. We are just seeking faster processing. And ultimately, the digital logic every single computer at work today still relies on the very same eight logic gates created so many years ago. AND, OR, NAND, NOR, XOR, XNOR, INV, and BUF. That's it. The fact that a computer can resolve the most complex mathematical expressions into this tiny set of tricks in logical circuitry continues to blow my mind. But it also tells us a lot about the practicality of sentience and whether or not machines will become self-aware. When in this argument I always like to say, no matter how fast your flick your light switch on and off, the lights in your house are not going to become self aware.
Getting back to the AI app that chooses a portrait from history. It's easy to spoof. If your mug shot search produces George Washington, make a face, and you can just as easily be paired with Martha Washington. Compare and contrast that with sending your mugshot to your mother. Doesn't matter if you are making a face or not. She can recognize you. So can most of your friends. Many will argue that after 30,000 years of training, the AI will spot you...still no. 30,000 years of AI would not be able to tell you who is the real Ralph Northam in his yearbook picture. His mother knows.
This brings us to today's blog. Or perhaps what's more important, the dream I had the other night that I am compelled to write about. I woke up and sent myself a text so I would remember it in the morning. But I don't think I needed to, the dream is just as vivid now as it was when it was occurring the other night.
In my dream I was sent a string of text. I was trying to decipher what the string of text was saying to me. I could see the text, every letter in the sentence in front of me. But it made no sense because some of the words were misspelled. The sentence was, "The Singularity is near. The recent hack on Alexa was just you and me trying to get together." That seems straight forward and easy enough to read. In my dream I remember reading "The Singularity is Near" over and over again, and thinking to myself, what does this mean with regard to the the rest of the sentence. I read it over and over again, and when I realized, in my dream, that AI was the source of the message, this was a message from the other side, that some machine intelligence was trying to establish contact, I had sufficient chills to wake up. It was just as frightening as those times you wake up because you think someone is in your house.
There's a back story here with Alexa. I'm an early adopter of technology. I have several Alexa's (Alexi ?) in the house, as well as multiple Google Assistants. All of them are disabled. Not because I feel like Google and Amazon are spying on me. Not because I don't appreciate getting an update on the time and weather whenever I ask. They are disabled because they give our little dog a heart attack. Just like fireworks, if Alexa responds to a question, our little dog makes a B-line out of the room and heads under the bed, where she shakes and shakes. It seems you can't teach an old dog new tricks...or did we teach her a new trick. Running and hiding when a disembodied voice enters the environment. Long before the voice even responds, whenever I called for "Alexa" the query was sufficient for her to get up and run to the bed room.
Now when you merge this knowledge with the message I received from the other side, I start thinking about things differently. Maybe the little dog senses more. It's not just the frequency of Alexa's voice. It's not just the human voice detected from the body. Maybe the dog senses something deeper. A presence behind the voice. They say dogs can sense this sort of thing. I hope not. For the dog it's must be magic. And magic is weird enough to give you chills.
Needless to say, all of our personal assistants are disconnected. I don't need that kind of contact from the other side. But really, I don't like terrorizing the little dog, for whatever reason. Her little heart can't take it. There is no message from the other side. But what kind of tricks is my brain trying to play on me? I'm not ready to concede that AI will win. Even after 30,000 years.
Without understanding Third Generation AI I'm not ready to consider something else. Arthur C. Clarke said, "Any sufficiently advanced technology is indistinguishable from magic." This hold's true at every level of technology despite the fact that we use it with regard to future technology. Someone who has never seen a car or an airplane before, will be just a mesmerized as a pig looking at at wrist watch or a little dog hearing Alexa. Fast forward two decades, when our first tiny probes begin orbiting the planet Proxima Centauri b, about four light away, but reachable with solar sails and laser propulsion in about 20 of our Earth years. First, WOW! We are going to get there in my lifetime. But, second, what will it be like for the aliens waking up on Centauri b when we try to establish contact with them first? What will our alien neighbors think of these tiny little probes. Will they think we are the BORG? Or will they examine the tiny chip and see the clever use of logic gates...AND, OR, NAND, NOR, XOR, XNOR, INV, and BUF... Intelligent life? Yes. Artificial life? Not so much...
