AI is so inescapable that it makes you wonder why you ever took a break in the first place. Not only are there countless chatbots and generative models, but the arrival and proliferation of this rapidly growing and ever-changing technology has become a constant, daily tinnitus for consumers.
The relentless strain of artificial intelligence is as catchy as Outkast's propulsive early 2000s song, “Hey Ya,” which helped perpetuate the myth that Polaroids had to be shaken to develop. And Polaroid photos are actually a lot like AI, needing no assistance to go from blank and blurry to sharp, clear, contrasty and defined.
The topic of AI follows me everywhere I go. Being a tech guy, everyone wants to talk to me about AI. This is understandable, since at work I often write about AI and experiment with different generative models to test their limits. But now the AI chatter has followed me home.
AI Appetizer
I was recently at a friend’s house, surrounded by at least half a dozen of my high school friends who I’ve known for literally decades. As we stood there sipping drinks (water for me) and munching on cheese and crispy snacks, one of my friends started casually asking me a question about AI.
As a banker, she understood how AI could be used in the workplace, but she was concerned with more personal, more fundamental questions: When will AI take over? When will we become too attached to robots? When will we love them?
We had a lengthy discussion on all these topics for about 30 minutes, where I explained that one of the reasons it's getting hard to assess the future of AI is that most of the AI model development, especially for the best AI chatbots, breaks the traditional 18-month development cycle of the technology and even Moore's Law, which states that the number of transistors on a chip doubles every two years, effectively doubling computing power.
But AI is amorphous and expansive. Try to find a beginning, a middle, and an end. AI is a vast collection of possibilities that stretch out in every direction.
In contrast, AI models are developed at a pace of three to six months, often exponentially improving their intelligence and capabilities by more than two-fold within that time frame.
The more we talked, the more uncomfortable I became. This was odd. Usually, I'm the one doing the arguing. I take difficult tech concepts and make them easy to understand for friends and family (and sometimes TV viewers). What seems scary and incomprehensible usually isn't. Technology, or most devices, are confusing when viewed as a whole. But when you break them down into parts and tasks, the picture becomes clearer, clearer, and easier to understand. Moreover, most technology has a beginning, middle, and end: a self-contained smartphone, computer, or even a VR headset. Software extends functionality but doesn't necessarily change it in a fundamental way.
But AI is amorphous and expansive. Try to find a beginning, a middle, and an end. AI is a vast collection of possibilities that stretch out in every direction.
Unjust Retribution
Another thing that worried me about the conversation was that I couldn’t reassure her that AI won’t take over certain jobs (basic writing, money and account management, customer service). I couldn’t find any argument that we’re not going to like AI.
In our last issue, we dove into the deep idea that humans are fundamentally complex biological machines, and the only difference between humans and robots with AI brains is sophistication. AI can only imitate human emotions so far, but who's to say that within a decade (or less) we won't have something that actually comes close?
I recalled that iRobot CEO Colin Angle had once said that C-3PO-esque robots were decades, if not more, away, and that we shouldn't expect to see them until 2050. But we hadn't spoken in a while, and perhaps the advent of generative AI has changed his mind. I told her about Future.AI and its new Future 02 robot, which, while it does walk poorly, has expressive torso and hands and has been given an imitation of personality by OpenAI's large language model.
Conversations with AI chatbots are commonplace these days, and in recent months, they have shown signs of being surprisingly human. When I told a friend that these were interactions that were once reserved for humans, she seemed intrigued and more than a little uneasy.
We agreed that the worry isn't necessarily about us — we'll be lumbering old by the time AI takes full control (we hope), it's our kids we should be worried about.
She recalled that she had once told her child that he or she could love anyone except a robot. “I was just kidding,” she began. “But I'm not kidding anymore,” I added.
I walked away from that discussion feeling a little shaken up.
As the night went on, my friend told another friend that we had talked about AI and that “Lance made me feel better.” The friend asked, “Really?”
“No,” she laughed, “No, he didn't.”
