“I THINK, THEREFORE I AM.”
I don’t know anything about 17th century philosopher Rene Descartes other than this quote. Cogito, ero sum is also the only thing I remember from high school Latin.
I do know that Descartes’ defining statement was more than a popular carriage bumper sticker (right up there with “Galileo Was Right.”) It was proof that the very act of thinking in itself was enough to confirm one’s existence. You might doubt your eyes, but you can’t deny your thoughts.
So what happens when we no longer need to think?
Stanley Druckenmiller, a billionaire investor (are there other kinds?) who wrote an op-ed for the Wall Street Journal that was 100 percent AI generated, defended himself by saying “Of course I used AI. I write everything using AI now for the same reason I use a calculator when I do math problems.”
He makes a fair point. I wouldn’t expect a billionaire investor to do long division any more than I would want my accountant to use an abacus when filing my 1040. I have a much different point of view on the right brain side of things, especially when it comes to prose and expressing lived experiences, but for now let me share some thoughts about thinking — not about oneself, as an insecure Descartes did, but for oneself.
When efficiency challenges human agency
Our team at PwC researches, builds, and executes AI and other emerging technology solutions. This includes synthetic reality avatars, physical AI and robotics, quantum computing and even space tech. We do it well, but what’s really impressive is how we see the future long before anyone else.
As you might imagine, this takes a lot of thinking, as well as the human intelligence required to parse the signals and make sense of that thinking. We take the necessary to time to connect the dots and look for connections that will lead us down new and interesting paths. But researching the future takes time, and we often need to place some bets and establish a POV quickly without sacrificing quality in the data or confidence in our judgments.
So our engineers and AI wizards, being the smart cookies that they are (and who are now cringing at being called cookies), built an AI research agent to help pull together the foundational knowledge and insights for a particular topic. Keep in mind, the agent isn’t just doing simple calculations or reading a few new stories or ArXiv papers. It’s not a Super Google compiling a bunch of already published research.
The agent thinks.
I know because I watched it. And when I say I watched it I mean I watched it for hours. I hung around to answer questions or provide additional guidance beyond my initial prompts, but mostly all I did, all I could do, all I had to do, was watch the machine think.
It was wild, like eavesdropping on someone’s inner monologue. Some really super insanely smart person’s inner monologue. The agent reasoned, second-guessed, sometimes changed its mind – and yes that was weird to write just now. Not only was I not in the loop, I didn’t even need to be in the room.
I HAVE SOME THOUGHTS.
Thinking through problems builds memory, and memory lays the foundation for intuition, expertise, and judgment. The more we offload our cognitive responsibilities, the more our memory erodes. As Druckenmiller would agree, why memorize complex equations when a calculator can do it for you perfectly, exactly, without fail?
That’s fine for math. Numbers don’t have biases or opinions. Math doesn’t have subjective interpretations of events or discoveries or hypotheses.
But people do. And that’s the point. Removing some friction is good – removing it all is dangerous.
Thinking requires rigor. If that sounds like a slog that’s because it is, but it’s that thinking and rigor up front that allows you to review, analyze, and better understand the AI’s output.
There’s still time for us to decide which cognitive tasks to hand over and which to hold onto for dear life. We can choose to watch the machine think or we can choose to think for ourselves.
Whether a thinking AI will ever be conscious is unanswerable and likely always will be, much like trying to answer “what does God look like,” or “what’s the sound of one hand clapping,” or “why do pickles exist?”
The bigger question is better left to the likes of Descartes.
Leave a Reply