AI and Economics Links
Rachel Lomasky on Hayek and LLMs; Keynes re-animated; Schumpeter responds; Joshua Gans on vibe research
AI is now at least partially a social science, a product of human complexity and profoundly shaped by human behavior and interaction. Social science, particularly Hayek's Complexity Theory, can provide essential frameworks for understanding the emergent systems we are building. More importantly, applying these insights can strengthen LLMs, enabling us to solve novel, challenging problems that neither humans nor computers could address alone.
Markus Brunnermeier uses AI to re-animate Keynes. Looking back on his famous predictions from 1930, Keynes says,
Material output per head has indeed multiplied. Yet, we have not taken the dividend entirely in leisure. On hours of labor, I was plainly mistaken. The 15-hour week, I envisaged, has not arrived.
But really, you have to watch the video to get the full effect.
Mark Latham brings back Schumpeter to criticize Keynes.
Keynes worries that we discover ways to economize on labor faster than we discover new uses for it. But this misunderstands the fundamental nature of capitalist development. The problem is not technological unemployment - it is the temporary dislocation inherent in creative destruction. This process is not a bug in the system; it is the essential fact about capitalism.
I cannot tell whether this is Latham giving the response or whether he asked an AI to respond as Schumpeter.
In my virtual wax museum you can choose either text or audio, but I did not do full multimedia the way that Brunnermeier’s resurrection of Keynes works.
The first reasoning LLM (ChatGPT-o1-pro) came out just over a year ago. It was the first model that opened up the possibility of using AI write at the start and through the process of a research project, all the way through to publication of results. I ran a little experiment in what I guess would now be called “vibe researching,” which took an idea I had long had (a fairly non-serious one) to see if I could, with the use of o1-pro produce a publishable paper in less than an hour. The answer turned out to be ‘yes,’ and the paper was published in Economics Letters as I recounted here.
He has been doing this for a year, and has more publications to show for it. Pointer from Tyler Cowen
substacks referenced above: @
What is gained by faking Keynes to comment on current affairs?
Before LLMs, we already knew that old predictions of '15-hours work weeks for most' have yet to come true.
More importantly, 'Keynes today' would not be 1930 Keynes. A "reanimated Keynes" (whatever that means!?) would be different. In his own life, he was a dynamic intellectual, adjusting his theories with the times.
Writing papers for publication in academic journals has been forever disrupted by Ai, of course. I have written with help and back and forth with my Ai two papers over the past two weeks. I have not sent them off for publication, because I find no value in that game, and make no mistake, it is a game.
Nonetheless, the papers could be published somewhere. Probably not in top journals, because we all know the super game that is all about. But who cares? The fact is that Ai assisted writing is already in full swing. The papers that get produced will be even more scarcely read than they already were before Ai. No big deal.
These days, if I want to research a question of any kind, be it economics or physics, I open up my Ai and start dialoguing with it. Before long, I know more about the question, or at least have had an intelligent conversation about it, than 99% of the people on the planet. Academia, both teaching and research, as we know it is dead as a door nail, is it not?