Monday , December 2 2024
Home / Real-World Economics Review / AJR, Nobel, and prompt engineering

AJR, Nobel, and prompt engineering

Summary:
From Peter Radford Well done AJR.  A prize deserved.  And remarkably little grumbling.  What’s wrong with that? In other news, my wife is deep into creating an artificial intelligence application.  One of the great challenges of getting AI to be useful is something called ‘prompt engineering’.   What have these two snippets of news have in common? The great thing about our better economists — the triumvirate we know affectionately as AJR being an example — is that they all seem to denigrate economics.  Politely of course.  Soto voce so to speak.  Or, in the case of AJR and Joe Stiglitz in books written for public consumption.  They key to this criticism is, I think, that once you are suitably credentialed — via tenure, status within the discipline, being well published etc — you can

Topics:
Peter Radford considers the following as important: , , , , ,

This could be interesting, too:

Lars Pålsson Syll writes Busting the ‘natural rate of unemployment’ myth

Lars Pålsson Syll writes Riksbankens oansvariga penningpolitik

Lars Pålsson Syll writes Debunking the ‘natural rate of interest’ myth

Merijn T. Knibbe writes The political economy of estimating productivity.

from Peter Radford

Well done AJR.  A prize deserved.  And remarkably little grumbling.  What’s wrong with that?

In other news, my wife is deep into creating an artificial intelligence application.  One of the great challenges of getting AI to be useful is something called ‘prompt engineering’.  

What have these two snippets of news have in common?

The great thing about our better economists — the triumvirate we know affectionately as AJR being an example — is that they all seem to denigrate economics.  Politely of course.  Soto voce so to speak.  Or, in the case of AJR and Joe Stiglitz in books written for public consumption. 

They key to this criticism is, I think, that once you are suitably credentialed — via tenure, status within the discipline, being well published etc — you can moan about the state of economics without damaging your ability to get those credentials.  It is as if you have to live two lives.  The first being acceptance of the illusion.  The second being the debunking of that illusion.  There are those, of course, who merge the two into one and persist in proselytizing the illusion even after they have achieved job security.  Those are the sort of people who keep telling the public that price gouging is a jolly good thing, and that there is no such thing as involuntary unemployment.

The problem with this is that it serves to diminish the public’s trust in economics.  If some of the great and good of the discipline are running around and debunking the nonsense put about by the likes of Milton Friedman and his loyalists, what are mere mortals to believe?

And, lets’ face it, economics needs to restore the public’s faith.  The disaster of the Great Recession lingers in everyone’s minds.  How come all those Nobel prize winners missed that one coming? And how is it that even with all the perfect information we all have, economists still can’t tell us what the economy will look like next year.  Accurately, that is.  

It just sounds odd.  Doesn’t it?

But, then again, I am a simpleton who thinks price gouging is an exploitative and nasty thing.  I lack the sophistication to understand just how wonderful it is.

The award of the Nobel prize to AJR sits in this context.

Their big idea seems to be that institutions matter.  I thought North won the same prize for the same idea decades ago.  Perhaps the prize committee forgot.  Or, perhaps they needed to remind us all that economies are multidimensional and have a lot more going on than those two dimensional charts we all remember from Econ 101 suggest.  Yes, I agree, AJR have pushed the boundary.  And, yes, they have hammered the point home.  

Well, not really.  I remain skeptical about the possibility of a wholesale re-invention of the textbooks just yet.  Inertia is a thing in academia.  Is it not?

You have to admit, though, that it is just a bit annoying for an outsider to plow through a big thick book written by a Nobel prize winning economist [or two in the case of AJR], and be told regularly that economics has major failings.  This and that are wrong, they say.  That idea misses the mark.  That other one is simply dangerous.  And this now, well, that’s just absurd.  It gives the impression that the house is not in order.  In fact, there seems to be quite a bit of disorder.  

And it isn’t as if the disagreements are contained to peripheral topics.  There seem to be some fairly significant disagreements over central bits of theory.  Oh dear.

One of my favorites is the whole general equilibrium fiasco.  What a silly idea.  But how tenaciously held it still is.  The amount of effort that goes in to clinging on to equilibrium is astonishing.  The word gets tossed around all over the place, and, in testimony to the impact economics has had in the public sphere, it has extraordinary currency in the business, financial, and related communities as well.  Experts talk about equilibrium all over the place.  They’ve never seen one of course.  That’s not the point.  But they all remember their economic classes.  Equilibrium is a big deal.  We all know that.  Don’t we?  Especially general equilibrium.  There’s that proof somewhere.  Right?  Arrow/Debreu, right?  

Absolutely. 

I am in the Arrow camp on that one.  What we fondly refer to as Arrow/Debreu is a definitive disproof of the existence of general equilibrium.  It is not a real thing.  It cannot be.  The conditions necessary for it do note exist.  Ergo, it does not exist.  

And yet the notion lingers on.  And on.  And on.

Nothing, it seems, gets forgotten.  Except for things that are ideologically unsound, but that’s another matter.

Back to AJR.

Institutions matter.  They are a big deal.  Culture matters too.  So do the social and other networks threaded through society.  Technology seems to have something to do with it.  There are all sorts of dimensions to take into account when thinking about an economy.  Nowadays the environment has become more important.  The list goes on.

It isn’t a simple trade-off.  It isn’t binary.  It isn’t two-dimensional.  Those simple charts we all grew up with are misleading.  They lead us to misguided end points.  Like price gouging is a jolly good thing.

Back to prompt engineering.  You thought I had forgotten didn’t you?

The problem with trying to get sensible responses from artificial intelligence is that the space of possibilities is riven through with dimensions that criss-cross through that space.  Ending up in a location that produces something of relevance and use is akin the proverbial needle in a haystack problem.  It’s a question of how you interrogate the system.  Ask the wrong first question and you are likely to end up with something that looks perfectly intelligent, but is irrelevant or, worse, misleading.  Each step on the journey refines the set of destinations.  Each question matters.  The sequence matters.  The breadth off potential space eliminated matters.  And so on.  This is difficult stuff.  It is clever stuff.  

Perhaps economics is like that.

Instead of starting with that final refinement, the cross hairs of supply and demand, we need to start with questions that set the stage better.  Which is what AJR do.  The institutional setting in which the next question can be asked shapes the possible responses.  It isn’t a two dimensional subject.  It isn’t about those cross hairs or simple trade-offs.  There are no two-dimensional frontiers to explore.  Economics needs to move beyond its eighteenth century roots and embrace the multidimensional nature of its subject.  The simplicity of its original questions have led us to responses that have been shown to be either incorrect or misleading.  Which is why so many top economists can describe so many problems in their public screeds.  Those problems are real.  And they exist because the initial questions economists posed led them into the wrong solution space.  The first prompt was wrong.  

Where we ended up looks intelligent.  Indeed it is.  It has all the trappings of a good response.  We have accepted it for decades.  It just isn’t very useful.  It has the aura of sensibility.  But it is analogous to an AI hallucination.  The reality it is useful within doesn’t actually exist.  Yes, that first prompt was wrong.  Or so it seems.

The prompt engineering of economics is something we need to pay more attention to.

Well done AJR!

Peter Radford
Peter Radford is publisher of The Radford Free Press, worked as an analyst for banks over fifteen years and has degrees from the London School of Economics and Harvard Business School.

Leave a Reply

Your email address will not be published. Required fields are marked *