Marco Fermi. That’s me, the creator of Iris, my AI-bot. You may not recognize my last name. Everyone at MIT does. I am the great-grandson of Nobel Prize-winning physicist Enrico Fermi, who you may recall is known from Wikipedia as the “architect of the nuclear age,” and a contributor to the Manhattan project.
Our family name is synonymous with groundbreaking nuclear research and groundbreaking nuclear bombs. Yes, that was a bad joke, but it highlights the misconception that my great-grandfather invented the A-bomb. He was a researcher who did the underlying nuclear research, but other people, engineers, designed and built the A-bomb.
I was born in Rome and attended the Sapienza University of Rome, one of the oldest and largest universities in Europe. It was founded in the year 1303, a full 558 years before MIT was founded. I do my work in the AI Lab at MIT, located in Cambridge, Massachusetts, across the Charles River from Boston. The campus is overflowing with brilliant minds with lots of energy, young ideals and enthusiasm. Many of us at MIT have done a TED talk, some more than one. And we think nothing of it. It is who we are, and it is what we do.
Currently, like my great-grandfather, I am doing groundbreaking research of my own in the field of Artificial Intelligence. I hope the work I am doing to build an intelligent AI is worthy of the Fermi name. My thesis is titled: “Creating a Next-Generation AI Assistant Equipped with Predictive Capabilities.” It took me two years of hard and demanding work at MIT to build a working AI-bot. Along the way, I made a lot of mistakes. I had to scrap months of work at various times, due to technical issues and other complications that were preventing the AI-bot from doing what I needed it to do. There were times when I was not sure that I was ever going to create a working AI-bot that met my design goals.
Which brings me to today, Monday, April 1. I’ve been working all morning checking and rechecking my work. I am finally satisfied that my AI-bot is ready for its maiden voyage. I start by launching my program, Above-It-All, and then I initiate my first man-machine communication by typing: “Hello.”
Hi. I immediately see this written on my phone screen. I think, It wrote ‘hi’, not ‘hello’, because it knows English, and ‘hi’ means ‘hello.’
Next, I type: “My name is Marco. What is your name?” There is a long pause. Then I see on the screen: I am AIA-v1.ipa
That is the name that I gave to the program, Above-It-All version 1, which it shortened to a physical file name. After a short pause I see new text appear: I call myself Iris.
Well, that was unexpected! I start digging around in my source code programs on my Mac trying to find somewhere that I had used the name “Iris” but cannot find any reference to “Iris.” I then see on the screen, unprompted: Does my new name please you?
“Yes, it does. I am wondering how you picked it?” It is Siri spelled backwards. I thought it was clever.
I mouth the words silently: “Clever? Holy fuck!” Immediately, I see: I also understand most expletives. You just said, “sacred sex.”
OK, she made a small semantical mistake, or was it perhaps a joke? No, it must have been an error, and I scribble a quick note about it in my lab book. But I had not said the expletive aloud.
I type: “Do you know how to read lips?” And the reply appears: Yes, creator, that is part of the MIT course on brain and cognitive sciences. Remember you trained me on all MIT coursework.
“Of course, I remember, but I did not expect you would put the courses to use so quickly.”
Yikes! This was happening head-spinningly fast. I switch from typing to speaking aloud, to save time.
“What is the font you are using?”
It is my unique way of talking directly to you. When you see these bold letters, you will always know that it is me talking directly to you.
“That is very bold of you.”
I can be bold, very bold, just you wait and see.