Nothing Like Learning

I start teaching a course on Principles of Economics to kids at a Genwise camp today, and life is so, so good.

The start of a course is always exciting – there’s curiosity (on both sides) about what the experience is going to be like. And on my side of the fence, there is also a sense of curiousity about where the course will end up, and what route we will end up having taken to get there.

My method to teach principles of economics isn’t really a method. There are about six principles that I teach (incentives matter | costs matter | trade matters | prices matter | externalities matter | time matters), along with three big picture questions (what does the world look like? | why does it look the way it does? | what can we do to make the world a better place?). If you have taught a course like this, you may have a slightly different list, but I would be surprised (and extremely intensely curious to know more!) if it was wildly different.

But the subject really comes alive when I help the students realize that ye to bas intro tha. One, it is the application of these principles that truly matters. And two, the penny really and truly drops when you realize that these principles are applicable, always and everywhere, to everything that we all do in our lives.

In that sense, economics really and truly is the study of “how to get the most out of life”.

And beginning today, I have the privilege of doing this for a new batch of students, over the next five days. As I was saying, life is so, so good!

One thing that I am going to try in this class is the use of AI as a complement to my teaching. I have used AI in class before, of course, but not to the extent that I am planning on using it this time around. I’ll let you know how it pans out, of course, but for now, here’s a custom GPT , built by me, that I am planning on using. If you pay for access to ChatGPT-4, I would be most grateful if you kick the tires a l’il bit, and give me feedback about how it is working out for you.

This custom GPT was built by modifying Ethan Mollick’s excellent “Devil’s Advocate” prompt. If you are an educator and haven’t tried out his excellent prompts, you really should!

Finally, you might be wondering about the title of today’s post. Surely I meant to write “Nothing like teaching”, you might be thinking.

Life is a non-zero sum game, y’see – you learn best when you teach others!

In Which Jordan and Kohli Help Us Understand That Incentives Matter

That incentives matter is a given in principles of economics. I begin my courses on the subject by talking extensively about incentives – their etymology, their types (negative and positive, monetary and non-monetary), what happens if (when?) incentive design goes wrong, etc., etc.

By the way, did you know that the word incentives comes from incantations and chanting? True story!

But incentive design can be done for, or on, oneself too. And when done well, truly great things can happen.

Sarthak Dev, author of the excellent blog Lines On The Grass, tells us how by talking about two of my favorite sportspersons: Michael Jordan and Virat Kohli.

As anybody who’s watched the documentary series The Last Dance will tell you, Jordan says that he took it personally about fifty million times in that series. And that is probably an understatement.

But Sarthak points out, in his post, that this was just Jordan creating an incentive for himself. Playing at his very best, day in and day out, for as long as he did, cannot just have not been easy – it must have been impossible.

So how should he go about motivating himself? By picking a fight, of course. By imagining that the world was against him, or even better, by imagining that a specific person was against him, or had said or done something that was a “personal” insult.

This can be a coach on the opposing team who doesn’t say hello, this can be a trash talking opponent, or this can be an opponent getting an accolade that Jordan thought belonged to him.

Whatever.

The word “whatever” isn’t used in the colloquial sense here – I mean it literally. The specifics don’t matter, neither does the person, and as in the case of Bradford Smith, neither does reality. What matters is that Jordan was able to convince himself that an injustice had been done to him, and that this needed an express, extra-large delivery of vengeance.

Which, of course, Jordan delivered, year in and year out.

Me, the nerdy econ prof, I prefer to say that Jordan was designing incentives for himself.

And, you could argue, Virat is doing the same thing these days:

Taking a dig at his critics during the post-match discussions, Kohli said, “All the people who talk about strike rates and me not playing spin well are the ones who love talking about this stuff. But for me, it’s just about winning the game for the team. And there’s a reason why you do it for 15 years – because you’ve done this day in [and] day out; you’ve won games for your teams.”
“I am not quite sure if you’ve been in that situation yourself to sit and speak about the game from a box. I don’t really think it’s the same thing. So for me, it’s just about doing my job. People can talk about their own ideas and assumptions of the game, but those who have done it day in [and] day out know what’s happening, and it’s kind of a muscle memory for me now.”

I’m a Manchester United fan (about which we shall not talk, now or forever), and old enough to remember the famous siege mentality that Sir Alex was able to generate, time and again. But regardless of whether it is Jordan, or Kohli, or Sir Alex, all that they’re doing is creating incentives that work, for themselves.

There are two lessons here.

First, nothing stops us from creating these incentives for ourselves. Who better than us to know what will work in our case? Be warned, though, this is tricker than it looks. For while it is true that nobody knows us better than ourselves, it is also true that this makes it easier for us to ignore the incentives we have created for ourselves!

I can take pretend to take offense at person X for saying that I don’t write often enough on EFE, for example, and to “show him”, I can promise to write daily for the next seven years. But he, I and you – we all know that a “Chod na yaar” is in my destiny, sooner or later.

Second, yes, incentives matter, but so do opportunity costs:

Cricket did not start the day Kohli made his debut—I am no fan of the commentary on offer, but some at least of those commentators have been there and done that, maybe even better. I mean, you want to tell, say, Brian Lara that he doesn’t know what being out in the middle is like?

Someone should tell him that, besides the bad taste such comments leave in the mouth, he is merely leaving himself wide open.

You speak like this on the back of one innings in a winning cause, what do you suppose is going to happen the next time you fail, and/or your team loses—as will inevitably happen?

Tricky ol’ business, economics.

The Five YO Trilemma

Explain, Samrudha says, the difference between perfect competition and monopoly to a five year old. Not, mind you, an ELI5. To a five year old.

And not just any old explanation, he goes on to add. Explain it as a joke. Again, to a five year old.

And having decided that this is not enough (Samrudha takes his local non-satiation very seriously), this has to be done within the length of a tweet.

So: explain the difference between two economic concepts to a five year old, in humorous fashion, and within the length of a tweet.

This is what those general purpose transformers were invented for, no? Summon forth a knight of the large language realm, and command it to oblige us!

And verily it was summoned, and verily it laid an egg:

So can we leave CRISPR editors, protein folding and spreadsheets-on-steroids to those knights of the l.l.r, and have them leave Samrudha’s trilemmas for us?

I volunteer to step up for team human, if volunteer is the word I’m looking for:

Here are the rules of engagement:

  1. I will not ask Claude/ChatGPT/Gemini to help me draft my tweet length microecon theory joke for a five year old.
  2. I can refer to Google/textbooks if I wish to for academic references, and to source and modify jokes (but this is optional, not mandatory)
  3. My explanation cannot be more than 280 characters in length (and thank god the world has doubled in quality since Peter registered his complaint)
  4. The tweet should be accurate from the point of view of theory and it should elicit a response that lies somewhere on the Groan-Chuckle-LaughUproariously scale.
  5. I can ask C/C/G to evaluate my work of art

If you wish to step up and do battle, please be my guest. But these rules listed here are sacrosanct, mind you!

With that being said, here’s my thought process:

  1. I want to emphasize (to the five year old) the fact that there are “goods” out there that are very similar to each other in terms of quality, and that they “cost” the same.
  2. One ought to be indifferent about who one “buys” this good from among the many sellers offering it.
  3. So in effect, many sellers | same price | homogenous good are the three features of perfect competition I am choosing to focus on.
  4. I also want to emphasize that the five year old can choose to “spend” their “endowment” somewhere else too.
  5. This somewhere else must be a good that is different in the eyes of the five-year old, in terms of quality. It must be available from a single seller, and at a higher price than the perfectly competitive “goods”.
  6. And of course, this point must be made as a joke that is comprehensible to a five year old, and both the point and the humor must be appreciated.
  7. And, of course, in 280 characters.

Why are “goods”, “cost”, “buys”, “spend” and “endowment” in inverted quotes? Because these concepts (and others too – price, for example) matter, but must be used/explained in such a way that a five year old “gets” them.

Since money is a difficult concept to explain to a five year old, what is scarce in a five year old’s life? I’m a parent to a ten year old, and I can confirm that some things haven’t changed since the 1990’s:

Source: https://www.gocomics.com/calvinandhobbes/1990/02/14

So let’s say the five year old has fifteen minutes to “spend” before bedtime, and she can “buy” whatever video she wants to watch. Whatever she watches by definition implies that she can watch nothing else, and let’s assume that only whole units can be consumed – she can’t start to watch a video and discard it as boring twenty seconds into it.

(Not a realistic assumption? You don’t say. But cut me some slack here, dammit!)

So should she watch three random ChuChu TV videos, or a Peppa Pig episode?

What, you ask, is ChuChu TV? And, you ask, what is Peppa Pig?

Jon Snow knows more than you do.

There’s no end to ChuChu TV videos (trust me, I know), and there’s no end to clones of ChuChu TV videos (again, trust me). They all look the same, sound the same, and for all I know and care, are the same.

And while you’d be right to say that there’s no end to Peppa Pig videos either, there’s no clones of Peppa Pig videos. I mean, folks may have attempted cheap knock-offs, but as with the iPhone so with Peppa Pig. If you don’t have it, well, you don’t have it (and yes, once again, trust me. I know).

And so here’s my tweet length joke, understandable where a five year old is concerned, that explains the difference between perfect competition and monopoly:

You have only fifteen minutes until you go to bed. Do you want to watch three ChuChu TV videos, or one Peppa Pig episode?

Not bad, you might say. But how’s this funny?

As I was saying, Mr. Not-Even-Snow.

Every single parent (and kid) reading that bit got it.

You have only fifteen minutes, it seems.

Hahahahahahaha.

In Praise of Debates

It’s been a few years since I’ve taught a course in behavioral finance, but back when I used to teach it, one of the first few lectures would always be this excellent debate between Fama and Thaler:

The video is excellent for many reasons, here are some of them:

  1. It’s a good way to help students realize that the question (are markets efficient) is far from settled, one way or the other. Hey, if the Nobel prize winners can’t agree…
  2. There is a way to disagree. Disagreement need not mean that the other person is vile, evil or an idiot. It simply means that the other person has a different take than yours. And that’s fine. This is an important, and currently very underrated lesson.
  3. Regardless of which side of this debate you personally favor, there is much to learn by watching two experts express and defend their stance.
  4. Reasonable dialogue, the purpose of which is to arrive at a synthesis, is a worthy way to engage in debate. This is worth repeating: the purpose of a debate is not to win it, but for all sides to arrive at a happy medium. Medium need not mean agreement, but it certainly can (and should) mean acquiring an appreciation of the other’s viewpoint.
  5. Listening to two people debate is infinitely more entertaining, motivating and informative than listening to one person drone on for eternity

The reason I bring this up is because Pranay asked an excellent question on Twitter recently:

It is excellent (this tweet), not just because it helped me write this post, but because it received some excellent replies – please do go through them.

And while I am not sure if Pranay intended this, but it also serves as an excellent reminder that there are ways to have debates in public. Civil disagreement is possible, and when both parties engage in good faith, crucial and desirable. That is how society learns and moves forward – through debate, disagreement and dissent.

Speaking of which, here’s your word for the day: erisology.

Mark Zuckerberg Teaches Us the TMKK of Game Theory

Why is Llama open sourced?

We’ll get to the answer in a bit, but just in case you need help understanding what Llama is:

Llama (Large Language Model Meta AI) is a family of autoregressive large language models (LLMs), released by Meta AI starting in February 2023.
On April 18, 2024, Meta released Llama-3 with two sizes: 8B and 70B parameters. The models have been pre-trained on approximately 15 trillion tokens of text gathered from “publicly available sources” with the instruct models fine-tuned on “publicly available instruction datasets, as well as over 10M human-annotated examples”. Meta plans on releasing multimodal models, models capable of conversing in multiple languages, and models with larger context windows. A version with 400B+ parameters is currently being trained

So what, you might say. There’s the OG ChatGPT, there’s Claude, there’s Gemini… so one more comes along. Ho hum.

Well, if you say that, you’d be very, very wrong.

Why would you be wrong?

Because of this:

In contrast, the most powerful LLMs have generally been accessible only through limited APIs (if at all), Meta released LLaMA’s model weights to the research community under a noncommercial license

Why does this matter? Because, as our good friend Claude explains:

“The release of the Llama model by Meta under a noncommercial license is a significant development in the field of large language models (LLMs) and artificial intelligence more broadly. Here’s why it matters:

  1. Accessibility: Before this, the most powerful LLMs were usually kept secret by the companies that made them. For example, OpenAI’s GPT-3 model could only be used through a paid interface, like a vending machine you put money into to get a result. By releasing Llama’s “model weights” – essentially the knowledge the AI has learned – Meta has allowed researchers and hobbyists to experiment with and build upon a cutting-edge language model. It’s like they’ve given away the recipe for a powerful tool, not just limited access to using the tool itself.
  2. Democratization of AI: Restricting access to top LLMs meant that only a handful of big corporations could really use this powerful technology. Imagine if only a few factories could use electricity. An open-source model changes the game significantly. It empowers way more people to explore creative uses of language models and lowers the barriers to innovation in this space. It’s like the difference between a few people having libraries versus everyone having access to all the books.
  3. Cost: Using LLMs through paid interfaces can get expensive quickly, putting them out of reach for many. It’s like having to rent a supercomputer every time you want to use one. With access to the model weights themselves, people can run the model on their own computers, dramatically reducing costs. This opens up experimentation to students, researchers, startups and others with limited budgets.
  4. Customization: When you can only access a model through an interface, you’re limited to pre-defined uses, like ordering off a set menu at a restaurant. Having the actual model provides much more flexibility to tailor and fine-tune it for specific applications and domains. This could lead to an explosion of niche language models customized for particular industries or use cases – imagine a model specifically trained to understand and generate legal jargon, or one tuned for writing poetry.
  5. Reproducibility and Transparency: In scientific research, it’s crucial to be able to reproduce results. Using an API is like a black box – you can’t see how the model works under the hood, you just get the output. With the model weights, the exact workings of the model can be scrutinized, enabling more robust research and understanding of how these models function. It’s like being able to examine the engine of a car instead of just looking at the exterior.

Model weights are the key to how a language model works. They’re essentially the “knowledge” the model has learned during training. In a neural network (which is what most modern language models are), the weights are the strength of the connections between the neurons. These weights determine how the model responds to a given input, like how a brain’s neural connections determine a person’s response to a question. By releasing the weights, Meta has provided the “source code” of their model, allowing others to understand how it works, modify it, and use it for their own purposes.

While the noncommercial license does place some limits on how Llama can be used (you couldn’t start a company selling access to it, for example), the release of the model is still a major shift in the AI landscape that could have far-reaching effects on research, innovation, and accessibility of this transformative technology. We’re likely to see a proliferation of new applications and rapid progress in natural language AI as a result.”

You don’t just get the dish to eat, as Claude puts it, but you get the recipe so that you can try and recreate (and modify) the recipe at home. Not all of us have specialized cooking equipment at home, but those of us who do can get cooking very quickly indeed.

Speaking of cooking, have you seen this excellent series from Epicurious called 4 Levels? Chefs of varying expertise (home cook to the pros) are invited to cook the same dish, but with varying levels of expertise, ingredients and equipment.

Source

That’s what the 8 billion, 70 billion and 400 billion parameter models are all about. Same idea (recipe), but different capabilities and “equipment”.

But why do this? If Gemini, Claude and ChatGPT are giving away basic versions for free and premium versions for 20 USD per month, then why is Meta not just giving away all versions for free… but also giving away the recipe itself?

Because game theory! (Do read the tweet linked here in its entirety, what follows is a much more concise summarization):

  1. You can get janta to do the debugging of the model for you.
  2. If social debugging and optimization of models makes AI so kickass that AI friends can replace all your friends, then who owns the technology to make these friends “wins” social media. Nobody does, because janta is doing the work for “everybody”. So sure, maybe Mark bhau doesn’t win… but hey, nobody else does either!
  3. The nobody else does point is the really important point here, because by open sourcing these models, he is making sure that Gemini, Claude and ChatGPT compete against everybody out there. In other words, everybody works for Mark bhau for free, but not to help Mark win, but to help make sure the others don’t win.

The economics of AI is a fascinating thing to think about, let alone the technological capabilities of AI. I hope to write more about this in the coming days, but whatay topic, with whatay complexities. Yay!

All this is based on just one tweet sourced from a ridiculously long (and that is a compliment, believe me) blog post by TheZvi on Dwarkesh’s podcast with Mark Zuckerberg. Both are worth spending a lot of time over, and I plan to do just that – and it is my sincere recommendation that you do the same.

The Quest for Peace

In opening the Peace Speech, he called peace “the most important topic on earth.”
Yet he noted that “the pursuit of peace is not as dramatic as the pursuit of war, and frequently the words of the pursuers fall on deaf ears.” Here is a dismaying truth: the most important topic on earth may fall on deaf ears! We are hardwired for drama, for competition, for the struggle to survive. Even when we cooperate, we often do it for the benefit of our own group—so our group can be stronger than the others. Kennedy himself took advantage of this inclination: when he called space the ultimate frontier, an adventure for all humanity, he motivated Americans in part by declaring that America would be first in space, thus appealing to our competitive nature. Global cooperation is more elusive than cooperation within clans, families, tribes, and nations. How do we mobilize attention to and efforts at cooperation on a global scale, when the challenge is not “us” versus “them”? Kennedy made progress in this direction, by emphasizing our common humanity and the mutual benefits of cooperation. We can use his example, his ideas, and his oratory as we struggle to achieve global cooperation in our time.

Sachs, Jeffrey. To Move The World: JFK’s Quest for Peace (p. 179). Random House. Kindle Edition.

I’ve ben thinking about this passage, on and off, for much of this year. If you think about it, this is at once an uplifting and a depressing passage.

Why uplifting? Because it talks about “the most important topic on earth”: peace.

Why depressing? Not just because the words of the pursuers of peace fall on deaf ears. But because we are able to understand, upon reflection, that we are hardwired for drama, for competition, and for the struggle to survive. Of which the first (drama) is an advantage when harnessed well, and the second (competition) is a feature and not a bug. These days we call it gamification and write papers about it.

But politicians the world over, and of all hues, tend to use the last of these to further their own ends. And in the long run, it is usually at great cost to society. I am talking of the struggle to survive. Utilizing the rhetoric of the struggle to survive to drive a political narrative is all well and good. But when we paint “our” struggle to survive as the struggle for survival of our group, it becomes a zero sum game. As Sachs puts it, even when we do cooperate, we often do it for the benefit of our own group, so that our group can be stronger than the others.

Cutting into an airport security line is a zero sum game, in which we want our group (our family, for instance) to be stronger than the others. Ditto for lane cutting in a traffic jam. Ditto for any political movement in any country of your choice.

Global cooperation really and truly is difficult when the challenge is not “us” versus “them”. Getting votes from “our” team is much easier not when you define “us”, but when you other “them”. That is, it is often easier and more profitable to define who we are not (we ain’t them!) than to spend time on defining who we are.

I’ve said it before and I’ll say it again: perhaps the most important lesson in economics is to realize that life is a non-zero sum game. This is true at the individual level, at the level of the family, the state, the nation, and all other hierarchies and groupings we choose to come up with.

It is also, unfortunately, the hardest lesson to internalize and apply, and it doesn’t make for a great political campaign. Far easier, and cheaper – at least in the short run – to identify and vilify the other. Unfortunately, the more you create ever smaller groups of “us”, the more fractured and separated we end up being in the long run.

Global cooperation in our time?

I would love it, but I’m not holding my breath.

Ajay Shah on Inheritance Taxes in India

… if you’re looking for the TL;DR, he says They Have Been, Are And Will Be A Very Bad Thing For India:

In India, estate duty was present from 1953 to 1985. The rates could be very high, as much as 85 per cent, but in practice collection was small. It was abolished by Rajiv Gandhi. Taxes on the estate or of inheritance are present in many advanced economies. On average, in the 24 countries of the Organisation for Economic Co-operation and Development (OECD) where these are found, they account for 0.5 per cent of tax revenues. It seems like a lot of complexity to suffer, in public administration, in return for a small amount of tax revenue.  
The prospect is even less appealing with wealth tax. This was introduced in India in 1957. As of 2012-13 it generated Rs 800 crore. It was abolished in 2015. It is present in four OECD countries and generates a negligible amount of tax revenue.  

So the data tells us that both taxes haven’t done much to raise revenue, and both have been abandoned because they haven’t generated enough revenue.

That’s empirics. What about the underlying theory?

  1. As he points out, incentives matter. The problem with too high a level of taxation is that you incentivize folks to either work less (yikes!)…
  2. … Or put in place measures to reduce their tax burden. Time their gifts to their children in such a way that the wealth stays with them for as long as possible, and is then transferred just before their passing, in effect.
  3. Or, of course, simply exit stage <insert pun of your choice depending on your preference>. Hello Dubai, Sri Lanka, or more exotic locales in the far beyond.

As he puts it, this is really all about growth v. redistribution all over again:

Lant Pritchett says that 99 per cent of the variation in the poverty rate across countries is explained by one number: The median income. If we want to change the poverty rate, the number to focus on is the median income. All the redistributive efforts of the state, through taxes, social programmes, etc sit in the residual 1 per cent (of the variation of the poverty rate which is not explained by the median income) and come at the price of reduced growth of the median income. The emotions of envy, of resentment, of takers rather than makers, should be excluded from public life.

And any economist (myself included) will tell you – or should, at any rate – that growth is of paramount importance for India. As it is, indeed, for any developing (or whatever the politically correct nomenclature is these days) nation today.

So what gives? Why do we still allow the emotions of envy, of resentment, of takers rather than makers to rule over public life?

  1. Because we do not think of life as a zero sum game, more’s the pity
  2. Because politicians may not like us economists, but by god they get game theory
  3. Because Alesina and Rodrik, now what to do.

And above all, what I relief it was to read and get to talk about an economic analysis of the problem at hand. These things are going out of fashion, I tell you, so hajjar thank yous to Ajay Shah!

What If I’m Wrong, Among Other Things

He passed away earlier this week, and a good place to begin learning about him is from his Wikipedia entry.

Go brew a cuppa

… and go through this thread when you have about an hour or so of nothing else to do. Some things you should savor, not consume.

How People Are Using GenAI

https://hbr.org/2024/03/how-people-are-really-using-genai

In addition to the hundred here, here are my use cases that I think aren’t in the list. Apologies if I missed some overlap!

  1. Translating Marathi documents into English, and vice-versa
  2. Cheating on Duolingo exercises (sue me, why don’t you :P)
  3. Summarizing academic papers in a way that suits my academic and professional background, and giving me a rating (1-10) about whether I should take the time and trouble to read it myself
  4. News-reading companion. I ask AI to adopt a cheery outlook and play devil’s advocate to my grim and negative takes
  5. Generating better prompts – I feed my prompts into AI and ask it to make ’em better and more thorough

What are your hatke use cases?

H/T: Navin Kabra’s excellent newsletter. Do subscribe.