Why learn when ChatGPT has all the answers?
Our washing machine started leaking the other day. We’d find puddles pooling near its legs at the end of a wash cycle. We couldn’t tell where or when the leaking would start. The puddles started getting bigger, but on occasion they would be small. The electrician we called over suggested a relatively expensive gasket change. The cynic in me, who has been right on occasion about this, assumed it was one of those unnecessary fixes for a smaller problem. But then what do I know about washing machines? Nothing. So we went ahead and made the change. And sure enough, it started leaking again. Not puddles, but trickles.
The cynic was incensed. After an irritated call with the electrician who professed ignorance it struck me I could try asking ChatGPT. It suggested a possible reason, which turned out to be the right one. It involved replacing an inexpensive drain hose. So it turned out that both the gasket and the pipe needed changing— or it wouldn’t have stopped leaking for a couple of days.
But that’s not the main point. ChatGPT plugged my ignorance and how! The mild trauma I have from recipe sites that make me scroll down half a kilometer before giving me something weakly useful tells me that using Google to diagnose this would have taken a while. ChatGPT bypasses all that and takes my half-baked layperson query and gives me the right answer. It’s digested all of humanity’s expertise and offers it to me “on tap.”
What’s that going to do to plumbers? Or electricians? Or computer programmers? Or any kind of work that requires codifiable knowledge? I use ChatGPT as shorthand here but it’s becoming evident that we’ll have a very wide array of LLMs fine-tuned for every domain. Why learn if knowledge and know-how are so effortlessly accessible?
I ask this not just as an individual but also as a parent wondering what systematic learning — education — should look like for our kids to navigate a world that’s changing at an unprecedented pace.
What should we learn, and why should we learn?
Running a microschool has made us realize that these questions inevitably telescope into questions about philosophy of life itself, and a theory of the self. We learn to acquire knowledge. What is this knowledge for? It depends on your view of society, self, and the soul. Our current form of education, in the way it is practiced not preached, assumes society is a series of zero-sum games. It assumes there are exclusive silos of knowledge — information and skills — we can acquire to excel at these games and more importantly outcompete others. This is why we clamor to get into IITs and Harvards and MITs. Winning those games, the thinking goes, sets us up well in this version and vision of society. Paulo Freire called this the “banking model of education” where we see students as empty vessels that need to be filled. Fortunately, there are other views. Each of these listed here is an alternative vision.
First, I must say that ChatGPT summaries and lists like these are wonderful time-savers. One doesn’t need to wade through a hundred turgid academic books with unreadable prose — Furthermore, it must be stated that it indeed is to no one’s advantage when the acquisition of knowledge is so heavily burdened by the unfortunate, yet essential imperative to signal to one’s fellow cohort of academics. — to know where to look or what to read. But summaries are lossy and we are faced with a different sort of question. Is the gaining of knowledge merely the acquisition of these summaries, or of the details these summaries sit at the top of?
Here, I find the very quotable Alfred North Whitehead’s vision of learning (tucked away at #19 out of 20 in the above list) to be the most prescient and in line with what we now know from neuroscience — even though his ideas are nearly a hundred years old. Whitehead believed that knowledge wasn’t a thing or some inert ideas one digested. It is a process by which we actively engage with the world and learn about it and our place in it and, ultimately, about ourselves.
“We think in generalities, but we live in details” — Alfred North Whitehead
The world, our society, our self, they are all complex nonlinear systems that appear to us as things. Nonlinear systems are often seemingly predictable, until they are not. Simpler systems, usually man-made, like washing machines are far more predictable and therefore codifiable. When you do this, it does that. What this breaks, that happens. These predictions work for everyone. But real-world complex systems often give different answers depending on who is asking, and how. This is a surprising aspect our traditional reductive view of science is not equipped for.
Before Jane Goodall discovered — after months of patient and painstaking observation — that chimpanzees used tools, the conventional view among primatologists and scientists was that tool-use was exclusive to humans. Because, well how could something so obviously sophisticated be within the reach of any non-human animals? Jane Goodall discusses how one scientist who had previously ventured into the forests to study chimps had done so with 22 porters. Talk about baggage blinding you to reality. Scientists at the time continued to be dismissive of Goodall’s discovery because it was anecdotal — mere detail — that did not fit the prevailing narrative stitched together from data. Flawed data collected with bias and blindness they didn’t even know they had.
Jane Goodall truly did help “redefine” mankind. Her work is an example that shows us that to learn more about the world and ourselves, we must be willing to question received wisdom and ask the right questions without assumptions and be receptive to the details. All learning is mediated by experiences and we are nothing but a constellation of experiences. Learning helps us define and shape our very contours, when we do it with our mind body and being.
We’ll find great uses for ChatGPT and AI and it’ll help usher in great change, hopefully for the positive. We will need to do this because, to paraphrase another Whitehead quote, civilization advances because of the innumerable number of little things we can do without thinking.
But as we do, the inevitable entropic march towards yet more complexity will continue. And as systems get more complex, the ones that decentralize well are the ones that persist. And decentralization requires that the thinking and learning are not outsourced entirely but done by each of us because life in all its details continues to stream in and the details are very different depending on where you stand.
So yes, we must continue to learn, especially by asking the right questions, because that’s how we discover and model the world, and find our self and ourselves.
