Glorb Blarba sat in the office watching his dad.
“Are we going soon?” he pleaded.
“Just a minute, Glorb. I’ve got to process the latest iteration and do a backup, and then we can get out of here.”
“You said that an hour ago! This is so boring. I don’t see how you can do it every day.”
Blorb Blarba stopped typing and looked at his son.
“Have you already finished your comic books?”
“I read them like three times, Dad.”
“Hm. I loved comic books when I was your age. It’s part of the reason I do what I do.”
“What? How?”
“I always loved the characters that had alter egos who had to use their mad science to figure things out.”
“Like “Iron-Rex?” asked Glorb.
“Exactly. He was a scientist first. I found it fascinating.”
“I guess I get that. What are you doing anyway?”
Blorb smiled. “I could tell you, but then I’d have to make you disappear.”
“Aw, come on. I never ask.”
“Come here, and I’ll show you.”
Glorb pulled his chair to the other side of his Dad’s desk.
He looked at the monitor.
“So here at Tricerawitz, we’re working on something very special. It’s something very important to me. I think the findings will be very important to everyone. And Robert Tricerawitz III, my boss, has entrusted me to run probably the greatest experiment of our generation.”
“What are you trying to figure out?” Glorb asked.
“Everything.”
“Everything?”
“Yes, everything. I’ll explain.”
“Have they taught you what machine learning is in school yet? Or deep learning. Or AI?”
“Yeah, it’s some kind of advanced programming thing.”
“Okay, okay, you’re more of an art student.” Blorb wiped his glasses on the edge of his shirt. “Let me break it down.”
Okay son. I’ll start at the beginning.
So computers are really strong now. They got this way in a crazy short period of time compared to how long we’ve been around on the planet.
We used to have simple computers with a limited number of transistors. But that number increased exponentially. Then we broke through into quantum computing making those capabilities seem like they were stone age. When I was born, we only just started using the internet broadly. There were hardly any mobile devices. Now the devices are built in.
“Wow, you’re old.”
Yes, yes, I am. Thank you.
Anyway, we started with simple programs and spreadsheets filled with data. Simple instructions that said if this data was input, then this would happen. Simple cause and effect. In effect, everything still is a more advanced iteration of that, but the amount of data that we can process and the amount of if-then instructions are nearly limitless.
Then when computers reached a certain capability, we started deep learning and big data. Where we would take many variables, sometimes disparate, and allow the computers to make correlations and findings that we would have never considered, or perhaps never had the time to consider.
“How do you mean?” Glorb said.
Well… say I was a realtor. And I was trying to sell a house. Previously we’d look at the average property value in the area, look at the size of the property, how new the house was, how big it was — both in area and in number of rooms and we’d create a value.
Deep learning would do this but it would also consider crime rate, proximity to an airport, train station, grocery store availability, and restaurants in the area, public school quality, proximity to a city, how good the local sports team is, weather, median, mean, and many many other things.
People used to think at a few combinations of those things to try and come up with statistical significance. Deep learning modeled how people think by passing relevant combinations of variables through neurons, which would then appropriately measure the weight of the variable.
“Which is where we get ‘neural network’ from, right?”
Correct. If you put several neurons together as a kind of layered system, you end up with a neural network. The amazing thing is that with new computing power we were able to create and weight more combinations of variables than ever thought possible. We were able to create associations and statistical correlations like never before.
Newer networks could adjust the weight and test new combinations on the go as well. It was revolutionary for data science.
Simultaneously, and in the same vein, we were creating machine learning algorithms to handle known data sets, and creating predictable outcomes. We would program all of the variables we wanted to know and train the program using data we already had. So that future information would render models based on well-known assumptions.
Later we would push into real artificial intelligence. Programs that were capable of reaching out and doing these things on their own terms, within the boundaries laid out by their programmers.
“That’s what we have.”
Correct. And they’re pervasive. Very quickly, people came to depend on them. Most jobs don’t exist anymore. Luckily, we have a benevolent government that curbed their use and allowed for them to solve some really big issues like healthcare, housing, and education. There was a crisis at first where AI started taking everyone’s jobs, but the economy still operated the same way — there was a massive influx of homelessness and uprising for around ten years. It was much worse then when we industrialized — because there truly were no jobs left. And the super rich who once controlled everything died in the Dinocrat space wars where each quadrillionaire took their social media zealots to the surface of Mars and had a giant pissing contest.
They’re currently up there now still, if they’re still alive. They didn’t like not feeling special, what with everyone else’s needs taken care of.
Now.
An interesting thing happened. Without jobs and with all this massive technology thinking of everything and taking care of all of our base needs. We started contemplating our own existence again. Velocimaslow’s hierarchy of needs were met for everyone.
“Velociwho?” Glorb asked.
Maybe you haven’t spoken about Velocimaslow yet. Basically, if your safety, shelter, sustenance, are all assured then all that is left is to answer the big existential questions. Like what am I here for? What is the purpose of any of this? Is there something bigger? Etc etc.
“I understand”
So here at Tricerawitz, we’re setting about answering those questions using the very technology we’ve built up.
Think about the best video game you’ve ever played.
“Easy. That’s the Legend of Zeldasaur.”
Well you know how big and realistic that game is? Well we’ve built kind of an environment similar to our own here in the world. It’s got the whole world in it. And it’s got trillions of AI’s.
We prerender a lot of it. But some things are not fully developed to save on computing power. For instance, they can’t typically get to the moon or mars but if they did, they’d find a partially rendered environment very monochromatic, because we didn’t waste the resources on a place they’d never get to.
“Just like in Skyrim! You can see Cryodilotopolis over the mountains”
Correct, even though it was released 240 years ago. They’re still remaking it for new systems.
Anyway, we built this environment and fed in all the data we know of ourselves. And similar to a normal machine learning path it takes several iterations to arrive at some kind of conclusion.
It’s probably going to take another 15 years to run the numbers and go through enough iterations.
But in that time we’ve processed what would happen to this planet a million times over from creation to its end. It takes nano seconds for us, but to the intelligence, it is a trillion trillion different ai’s being created, experiencing what they do throughout their life, given the experiences of the time, and then we harvest the information after it returns to the core. We wipe them clean and send them through again.
“Have you found anything interesting?”
Oh, so many things. Perhaps the most interesting is that about half the time we do not survive and are replaced by the future generations of evolved mammals.
“No way!”
Yes, I know. Preposterous given their state now.
But either way, no matter who evolves, if they don’t destroy themselves once they discover the atom, then they also discover computers and eventually discover artificial intelligence as well.
“And?” Glorb was actually interested now.
In this last iteration, take a look. They went through a very similar struggle and are currently beginning to toy around with the idea of using machine learning to determine the meaning of life by setting up a giant simulation!
“Did they?”
Well, not in this case. But this is the farthest I’ve seen them go. They usually destroy themselves within a few zeptoseconds…about twenty years for them.
But I will say that if it happened once. It’s likely to happen again. Maybe they’ll find the meaning before we do, ay?
“I’d hate to ruin things for them by telling them they’re just there so we can gather information.”
It’s a noble cause son. Otherwise we dinosaurs will likely never understand our purpose.
Glorb looked at his dad. He was proud.
“Wow. I didn’t know you were doing that.”
Blorb smiled. He hit escape on his keyboard.
“Well son, I think it’s time we go home. I’m all finished up. Maybe we can stop by the comic book store on the way home.”
Glorb jumped up. All this talk of AI and machine learning gone in an instant.
