“A Series of Small Apocalypses” Out Now in Liberties

Plus, The Complete ChatGPT-4 “I Can’t Believe It’s Not Butter” Transcripts

I have a long essay out in the current issue of Liberties. Please go read it, and subscribe to that august publication too. It develops some of my ideas about the “real” threat of artificial intelligence. You’ve seen many of the points I make there in this space already, but naturally there I make the case in a more polished and less digressive way. Here is an excerpt:

A delicious irony, that this precipitous rise of the machines might be what it finally takes to break us free of the machine metaphor of the human mind, and to open up the possibility for a vision of education based fundamentally on experience: something that we have no evidence that machines, for their part, are capable of having. We have no falsifiable evidence that human beings are capable of having it either, but humanistic education has always set out from the assumption that they do. In this era of STEM imperialism, when the methods and the requirements of the natural sciences increasingly dominate the humanities, and increasingly require of humanists that they justify their work in terms imposed on them by the STEM fields, we are indeed in a moment of severe crisis. But no crisis, short of total apocalypse, ever comes without a glimmer of hope.

Go read the whole thing now!

Meanwhile, I’m working under several stressful deadlines now, and do not think I will be able to deliver a new essay this week. So I’m going to do what everyone else is doing these days, and turn to ChatGPT to do my work for me. Some of you have already read, in Substack Notes, an excerpt of a dialogue I had with that trusty AI companion on the subject of I Can’t Believe It’s Not Butter. What you did not realize, when you saw that, is that it is just a very tiny part of a very, very long exchange.

I am going to withhold myself from offering reflections, for now, on the lessons that might be drawn from this, on what I’ve learned from the exchange, or what I think it reveals about the prospect for “general AI” in the near future. I invite all of you, dear subscribers, to read the text and to let me know your own thoughts. Are you impressed? Are you unimpressed? Is our commercially available ChatGPT getting better at certain operations? Which ones? How far does it still have to go?

I’ll just say that as regards its ability to simulate wit, at present AI reminds me of nothing so much as a German academic trying to land a joke in English at an academic conference. Which is to say at least in this crucial domain I think it still has a long, long way to go.

But do see for yourself…