The Solow Model and China

If you don’t know what the Solow model is, here is a great place to get started:

There are 11 videos in that series, and if you can spare the time, please watch all of them. Just two a day (they’re not more than 5 minutes each), and you’ll be done come the weekend.

But in effect, here is what the Solow model says:

  1. Output for a nation is a function of three (actually four) things:
    1. Capital (K): Buidings, ports, dams… infrastructure, basically.
    2. Education Augmented Labor (eL): The amount of hours that a person is able to put in to their work, but with the built in assumption that an educated person is likely to be more productive than a person without education.
    3. Ideas: Read the paragraph below to get a sense of what this means in practice.

Think about this blogpost that you are reading. I wrote it using my laptop, which is my capital. I will spend about an hour (that’s my plan, I’ll update you towards the end of this post about how well it worked out) writing it, and that’s the labor that I’ll be putting into this post. The fact that I have been “educated” in economics should mean that this post will be easier to write for me than, say, a gardener. The gardener could have written this post as well, of course, but it’s safe to assume that she would first have had to learn about the Solow model, and that, presumably, would have taken longer.

So that’s K and eL where the output (this blogpost) is concerned. But now think about it this way: what if another person, with a similar level of economics education as mine were to write this blogpost instead of me? Would that person have chosen this video, and these paragraphs to explain the Solow model? Maybe they would have recommended some other video, or some other podcast, or chosen to share details of an online textbook in which the Solow model is explained. That’s one way to think about ideas.

And so when you combine the capital (the laptop), the labor (the time I spend on this blogpost, given my education levels) and the ideas (what I choose to put into this blog post, and how), you get the output you’re reading right now.

What if I double the capital? Will the blogpost be done in half the time? Say I have an external monitor attached to my laptop – will two screens mean finishing the blogpost in half the time? It will save some time, but not by a factor of two, surely. Trust me, I have tried.

What if I double the labor? Hire an assistant to write this blogpost with me? The way I work, trust me, it will probably take longer! What if I go get a post-doc, to augment my education? Will that save me time? The hysterical laughter you hear in the background is the response of any PhD/post-doc student anywhere in the world, and that sound means a loud and resounding no.

In a sense, the Solow model asks these and related questions, and answers them using some graphs and equations. Except, of course, the Solow model does it for not one guy writing one blog, but for an entire nation at a time. There is no sense in me explaining the whole model over here, for it would be a case of me reinventing what is already a very good wheel. Please watch the videos.


But the Solow model is a remarkably useful way to get a handle on the long run growth prospects of a country. Is India likely to grow in the future? Well, is it going to add to its capital stock? Yes. Is it going to augment it’s stock of education augmented labor? Yes. Is it likely to produce more ideas than it is right now? Yes. And so the growth prospects for India look reasonably good.

Of course, there is more to the Solow model. All of this holds true given a strong and stable political system, well established rules of law, and strong and capable institutions. But so long as you believe that these are likely to continue to be so in the Indian case, you should be bullish on India.

What about, say, Japan? It has a capital stock that is more in need of replacement than new construction ( a feature of the Solow model that we have not discussed here, called depreciation), so it is unlikely that it will grow its capital stock too much. Here’s an example of what I mean. What about it’s stock of education augmented labor? Well, the news ain’t very good. Ideas? Trending upwards, but not by much. So if I had to bet on which country would grow more over the next twenty years, I would bet on India, not Japan.

Bear in mind that this is a model, and like all models, it is an imprecise abstraction of reality. So it is possible that at the end of the twenty year period, we find out that I am completely wrong. But if you think the Solow Model is a reasonably good model, you ought to bet the way I did.


So what about China?

Well, now, that’s a whole different story, and one that Noah Smith talks about in a recent blog post. Long story short, he doesn’t think China’s growth prospects are that great.

But the story is a little more complicated than that. The Solow model is a good model, sure, but it’s not as if the Chinese authorities/experts aren’t aware of the problem. And in his blog post, Noah looks at arguments put forth by two people who know a thing or two about China, and analyzes them critically.

The first argument is that sure, China’s demographics are on a downward trend, but what if we raised the retirement age for Chinese workers? Would that not solve the problem? Noah says no, probably not, because firms made of exclusively old folks isn’t necessarily a good idea. I wholeheartedly agree.

What about adding to China’s urbanization, and therefore its infrastructure? After all, China’s urbanization rate is “only” 64%. The inverted quotes around only in the previous sentence is because we, in India, are officially at 31%, but as in the case of China, it very much is a function of how you define urbanization. But similarly, in China, the urbanization rate is actually way more than 64%, and the Lewis turning point has already taken place in China, or will do so any moment.

And about ideas, well, China is an even more complicated story. Noah makes the point that China’s industrial policy is essentially a one-man army that is trying something that has never been tried before, and Noah is betting on it not quite working out. And given the events of the last year and a half or so, it is hard to disagree.

And so the Solow Model would probably tell you that China is unlikely to grow as fast in the near future as it did in the recent past, and even if you take into account potential adjustments, it likely will still be the case that China’s growth rate will start to plateau.


Please, read the entire post by Noah. But if you are a student of economics who has not yet met the Solow Model, begin there, and then get on to Noah’s post – your mileage will increase considerably.

Have we become uniquely stupid?

For those of you who have read the essay, the title of today’s blogpost is a dead giveaway: I am referring to Jonathan Haidt’s essay in the Atlantic, titled “Why The Past 10 Years Of American Life Have Been Uniquely Stupid“. The subtitle is equally depressing: It’s Not Just a Phase.

It’s been clear for quite a while now that red America and blue America are becoming like two different countries claiming the same territory, with two different versions of the Constitution, economics, and American history. But Babel is not a story about tribalism; it’s a story about the fragmentation of everything. It’s about the shattering of all that had seemed solid, the scattering of people who had been a community. It’s a metaphor for what is happening not only between red and blue, but within the left and within the right, as well as within universities, companies, professional associations, museums, and even families.
Babel is a metaphor for what some forms of social media have done to nearly all of the groups and institutions most important to the country’s future—and to us as a people. How did this happen? And what does it portend for American life?

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

The essay is a lengthy read, but a rewarding one. Jonathan Haidt takes us through the evolution of the internet, with the emphasis on the social aspect really beginning to take off post 2010 or so, and gives us a book to read that goes on my to-read list: Nonzero: History, Evolution and Human Cooperation.

The next section is where the story really picks up, for we are introduced to the “villains” of the piece: the Like, Share and Retweet buttons. It’s not the buttons themselves that are to blame, of course, much like the atom not being at fault for the atom bomb. It’s what we have done with the Like, Share and Retweet buttons that is the problem:

By 2013, social media had become a new game, with dynamics unlike those in 2008. If you were skillful or lucky, you might create a post that would “go viral” and make you “internet famous” for a few days. If you blundered, you could find yourself buried in hateful comments. Your posts rode to fame or ignominy based on the clicks of thousands of strangers, and you in turn contributed thousands of clicks to the game.

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

Goodhart’s Law is massively underrated. Rather than optimizing for the quality of the content of one’s creation, we optimize for it’s virality. The virality ought to be a function of the quality, but we’ve skipped the intermediate step, with consequences that have become manifest and deep-rooted. Or as Jonathan Haidt puts it, “these platforms were almost perfectly designed to bring out our most moralistic and least reflective selves”.

He then goes on to quote from Madison’s Federalist No. 10 on the innate human proclivity towards “faction”.
I have watched “The Last Dance” on Netflix more times than I should have, but this reminds me of Michael Wilbon talking about how everybody in Chicago hated the Pistons (around the 28 minute mark in episode 4, if you’re interested). He repeatedly involves the phrase “this was personal”, and that’s one way to understand what factionalism means. Tribalism in sports, but elsewhere too, is the kind of factionalism you want to think about in this context, and you might also benefit from reading the transcript of Ezra Klein’s conversation with Tyler Cowen:

https://conversationswithtyler.com/episodes/ezra-klein-2/

Factionalism (or tribalism. I’m not sure if the two mean exactly the same thing in an academic sense, but I am using them interchangeably here) hasn’t necessarily gone down, but we seem to have found new things to be “tribal” about.

As I understand it, Haidt is making the point that our tribalism when it comes to politics is now more deep-rooted than ever, but is also more trivial than ever before. Which politician is wearing what kind of clothes for which occasion excites more debate online than substantive issues that warrant more debate. Or as I prefer to put it, our agreement with stated positions and policies is these days a function of who said it, rather than what has been said. Such tribal loyalty when it comes to close friends is one thing, although even that has its limits, but fealty of such an extreme nature when it comes to political discourse ought to worry most of us.

And as an aside, the last question that Tyler Cowen asks in that extract above is a question to which I don’t have a great answer. I agree with the point in his question, but like him, wonder about the underlying cause.


An extract twice removed now:

The digital revolution has shattered that mirror, and now the public inhabits those broken pieces of glass. So the public isn’t one thing; it’s highly fragmented, and it’s basically mutually hostile. It’s mostly people yelling at each other and living in bubbles of one sort or another.

https://www.vox.com/policy-and-politics/2019/12/26/21004797/2010s-review-a-decade-of-revolt-martin-gurri

Amit Varma made a very similar point in a recent podcast with Shruti Kapila recently, in which he pointed out that social media has, in effect, decentralized the news (I’m quoting from memory here, so please forgive me if I’ve got the exact wording wrong). Amit Varma says that this is on balance a good thing, but with some negative consequences. Jonathan Haidt disagrees:

Mark Zuckerberg may not have wished for any of that. But by rewiring everything in a headlong rush for growth—with a naïve conception of human psychology, little understanding of the intricacy of institutions, and no concern for external costs imposed on society—Facebook, Twitter, YouTube, and a few other large platforms unwittingly dissolved the mortar of trust, belief in institutions, and shared stories that had held a large and diverse secular democracy together.

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

Where do I fall on this Haidt-Verma spectrum? Closer towards the Haidt end, I’d say, but I do have to remind myself that I have written this and you are reading it, so maybe decentralization isn’t all that bad? But that’s as far as I’m willing to go – on balance, I find myself closer to Haidt’s position, at least for the moment.


But the enhanced virality of social media thereafter made it more hazardous to be seen fraternizing with the enemy or even failing to attack the enemy with sufficient vigor. On the right, the term RINO (Republican in Name Only) was superseded in 2015 by the more contemptuous term cuckservative, popularized on Twitter by Trump supporters. On the left, social media launched callout culture in the years after 2012, with transformative effects on university life and later on politics and culture throughout the English-speaking world.

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

Haidt is writing this from an American perspective, for an American audience. But we in India have our own share of names for The Other, don’t we? It’s not just the fact that we have relatively trivial tribalism in areas as important as political discourse, but the fact that the discourse itself is not just trivial, but downright nasty. And the nastier it gets, the higher the support from your own side!


I’ll skip talking about a couple of sections from Haidt’s essay, not because they’re not important, but because they aren’t directly relevant to us here in India. But the subtitle of his essay gets an entire section, where he speaks about how things are likely to get much worse in the years (months) to come:

in a 2018 interview, Steve Bannon, the former adviser to Donald Trump, said that the way to deal with the media is “to flood the zone with shit.” He was describing the “firehose of falsehood” tactic pioneered by Russian disinformation programs to keep Americans confused, disoriented, and angry. But back then, in 2018, there was an upper limit to the amount of shit available, because all of it had to be created by a person (other than some low-quality stuff produced by bots).
Now, however, artificial intelligence is close to enabling the limitless spread of highly believable disinformation. The AI program GPT-3 is already so good that you can give it a topic and a tone and it will spit out as many essays as you like, typically with perfect grammar and a surprising level of coherence. In a year or two, when the program is upgraded to GPT-4, it will become far more capable. In a 2020 essay titled “The Supply of Disinformation Will Soon Be Infinite,” Renée DiResta, the research manager at the Stanford Internet Observatory, explained that spreading falsehoods—whether through text, images, or deep-fake videos—will quickly become inconceivably easy. (She co-wrote the essay with GPT-3.)

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

Speaking of the amount of shit that had to be created by a person, read this article written by Samanth Subramanian in February 2017.


So what might be done? Jonathan Haidt has a three-pronged solution:

What changes are needed? Redesigning democracy for the digital age is far beyond my abilities, but I can suggest three categories of reforms––three goals that must be achieved if democracy is to remain viable in the post-Babel era. We must harden democratic institutions so that they can withstand chronic anger and mistrust, reform social media so that it becomes less socially corrosive, and better prepare the next generation for democratic citizenship in this new age.

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

He outlines the steps involved in each of these, and if you haven’t already, I would encourage you to go read the entire essay, and these outlines in particular. I find myself to be in broad agreement with both the suggestions as well as how they might be implemented, but also worry about whether we have the political and social will to actually do so.


Finally, a coda of sorts:

The most pervasive obstacle to good thinking is confirmation bias, which refers to the human tendency to search only for evidence that confirms our preferred beliefs. Even before the advent of social media, search engines were supercharging confirmation bias, making it far easier for people to find evidence for absurd beliefs and conspiracy theories, such as that the Earth is flat and that the U.S. government staged the 9/11 attacks. But social media made things much worse.

https://www.theatlantic.com/magazine/archive/2022/05/social-media-democracy-trust-babel/629369/

And I would feel very bad if you, the reader, were to read either my post or Haidt’s essay in order to confirm your already existing fears about the ill-effects of social media. And so I urge you to read this column by Tyler Cowen next:

Calling something “extremist” is not an effective critique. It’s a sign that the speaker or writer either doesn’t want to take the trouble to make a real argument, or is hoping to win the debate through rhetoric or Twitter pressure rather than logic. It’s also a bad sign when critics stress how social media have fed and encouraged “extremism.”


What the U.S. needs is more consideration of more extreme ideas. If you see someone inveighing against “extremism” or “extremist ideas,” beware: That is itself an extreme position. True moderation lies in calm and reasoned debate.

https://www.bloomberg.com/opinion/articles/2022-05-06/extremist-ideas-are-not-always-bad-and-are-often-popular

My take on this essay? I think Tyler is saying that we shouldn’t be throwing the baby out with the bathwater. Social media has done two things: made it easier to spread “extreme” ideas, and made it much more likely that we will react with extreme prejudice and nastiness to these ideas.

The first of these is A Very Good Thing and the second of these is a Very Bad Thing. But we would do well to hold on to the first, rather than abandon both.

How? Ah, now if only we had some extreme ideas about that.

Feynman on Taking The World From Another Point of View

If you have not seen this series before, you’re going to enjoy your Sunday morning. Remember, it is a four part series, I have linked to only the first one – but trust the YouTube algorithm to help you out with this after you’re done with the first!

Futurology from 1967

Did no work of science fiction/futurology anticipate miniaturization? Genuine question.

About Presentations

A student wrote in asking about my ideas about presentation design for a project that she I and are working on together. Here is advice I have found useful, based on what I’ve read/seen online, and for having delivered and sat through presentations:

  1. Be clear about how the presentation is going to be used. Is it meant to be read by participants as a hand-out or on their screens, without you or somebody else being around to deliver a talk? That means lots more details, more notes, and a much lengthier and text heavy presentation. My sincere advice would be to not design a presentation, but to use a document (Google Docs/MS Word/whatever you prefer) instead.
  2. But if the presentation is a complement to what you – the speaker – are going to say, use the 10-20-30 rule.
    “It’s quite simple: a PowerPoint presentation should have ten slides, last no more than twenty minutes, and contain no font smaller than thirty points.”
  3. You’re taking up the audience’s time: value it. If you have thirty people listening to you for thirty minutes, that’s fifteen hours that could have been spent doing something else. Your presentation should be worth that time, you have a huge responsibility on your shoulders. Not enough people appreciate this, unless they happen to be in the audience.
  4. Ruthlessly edit the number of words on your slide. The “30” part of the 10-20-30 rule is a heuristic, and the idea behind it is to force you to use lesser words.
  5. The words that do remain on your slide should be “keywords”, and should be words that you want your audience to remember after the talk is over. If this is not true, then why do you have that word on your slide in the first place?
  6. If you’re presenting a chart, or a table or an infographic, make sure the corresponding title answers the question “So What?”. If your chart shows that sales have gone up in the last four quarters, don’t title the chart “Sales Have Gone Up In The Last Four Quarters.”
    Does that mean you need to hire more workers? Increase inventory? Increase shift timings? Each of these (and so many more) are the “so what’s”. Make the most important of these the title of the chart.
  7. Don’t use the default color template that PowerPoint (or any other software) gives you. Take the time and trouble to figure out how to change the color template, and use one that is appropriate for your presentation. It helps make your presentation more memorable. But also note that simplicity is underrated!
  8. Check, double check and triple check for spelling mistakes. (I’m being a hypocrite right now, because I discover typos in my older blogposts all the time, and it kills me). For presentations, I usually add a thick black diagonal line to each slide, and only remove it after I know that I have double checked each word and element on that slide.
  9. For truly important presentations, have somebody else do the same thing after you’ve done it yourself. A fresh pair of eyes really helps!
  10. Always be prepared to deliver a presentation without the corresponding PPT. Yes, you may have back-ups, but there will be the occasional time when nothing has worked and everything has failed… but the show must go on. Be prepared!

The meta-epistemology of the rate hike

Soon after I started blogging, Tyler Cowen joked, “You’re not really a blogger.” His point: Unlike most of the competition, I wasn’t reacting to the latest news or whatever’s hot. My goal as a blogger has always been to write think-pieces that stand the test of time.

https://www.econlib.org/a-fond-farewell-to-econlog/

I don’t know about standing the test of time where posts on EFE are concerned, but my approach to blogging is very similar: I prefer to not write about events immediately after they’ve occurred. This for a variety of reasons, not least of which is the fact that I’m lazy, and reading a lot of stuff at very short notice is something I would rather not do.

Another reason is that the very best pieces on any event usually take time to bubble up in my feed, and waiting therefore makes sense.

By the way, if you aren’t yet subscribed to Bryan Caplan’s new blog, please do!


But that being said, let’s talk about yesterday’s rate hike.

One of the pieces that I enjoyed writing last year was on the concept of meta-epistemology, after reading a post about it by Zeynep Tufekci.

I’m going to post a screenshot rather than an extract, because the formatting of the post helps:

https://econforeverybody.com/2021/02/05/zeynep-tufekci-on-metaepistomology/

Honest question: does this apply to the Reserve Bank of India as well?

Is it the case that the cost of downplaying inflation as a major problem now exceed the benefits of doing so? Have the incentives flipped for the RBI? If so, on what basis? Is there a sense, based on preliminary data, that inflation is a problem that can no longer be ignored?

And if so, how should we be interpreting not just the fact that rates have been raised, but the manner and the timing of the raise? In other words, are there two messages being sent out by the RBI: the message itself, and the implicit message encoded in the timing of the message?

And have (or will) the markets internalize this message, and if yes, what is to follow?


Learning about inflation, monetary policy, and the efficient market hypothesis via textbooks is less than half of the story. Take your view/model of how the world works to the world itself, and update your model as the years roll by.

Fun, exhilarating and occasionally nerve-wracking.

But it is the best way to learn.

Happy Birthday to Kevin Kelly

70th birthday that too!

Who is Kevin Kelly, you ask? Lots of ways to begin, but my favorite learning from Kevin Kelly (so far) has been the idea of 1000 true fans:

To be a successful creator you don’t need millions. You don’t need millions of dollars or millions of customers, millions of clients or millions of fans. To make a living as a craftsperson, photographer, musician, designer, author, animator, app maker, entrepreneur, or inventor you need only thousands of true fans.
A true fan is defined as a fan that will buy anything you produce. These diehard fans will drive 200 miles to see you sing; they will buy the hardback and paperback and audible versions of your book; they will purchase your next figurine sight unseen; they will pay for the “best-of” DVD version of your free youtube channel; they will come to your chef’s table once a month. If you have roughly a thousand of true fans like this (also known as super fans), you can make a living — if you are content to make a living but not a fortune.

https://kk.org/thetechnium/1000-true-fans/

I cannot for the life of me remember where I read about 1000 true fans first, but it most likely was via Tim Ferriss. (As an aside, Kevin Kelly has advice about this as well!) The extract above is an assertion, and if your reaction is along the lines of “but why is this assertion true?” – and I hope that is the case! – you will want to read the rest of the essay. It’s got spin-offs too, this essay, which only drives up my opinion of the original.

But Kevin Kelly is a person who you should spend time learning more about. Start with his Wikipedia page, listen to his multiple episodes with Russ Roberts over on EconTalk, visit the Cool Tools section on his website, subscribe to his related newsletter, listen to his podcasts with Tim Ferriss, and as a bonus, listen to Tyler Cowen’s podcast with Stewart Brand. And read his books, of course.

Long story short, he is a person worth knowing about, and trust me when I say we’ve only scratched the surface, if that. But today, I wanted to point you to his birthday gift to all of us, a lovely set of 103 observations that he has called “103 Bits of Advice I Wish I Had Known“. It goes without saying that all 103 are worth a ponder, but I’ll list here ten that especially resonated with me right now:

  1. About 99% of the time, the right time is right now.
  2. Anything you say before the word “but” does not count.
  3. When you forgive others, they may not notice, but you will heal. Forgiveness is not something we do for others; it is a gift to ourselves.
  4. When you lead, your real job is to create more leaders, not more followers.
  5. It is the duty of a student to get everything out of a teacher, and the duty of a teacher to get everything out of a student.
  6. Productivity is often a distraction. Don’t aim for better ways to get through your tasks as quickly as possible, rather aim for better tasks that you never want to stop doing.
  7. The consistency of your endeavors (exercise, companionship, work) is more important than the quantity. Nothing beats small things done every day, which is way more important than what you do occasionally.
  8. Half the skill of being educated is learning what you can ignore.
  9. When you have some success, the feeling of being an imposter can be real. Who am I fooling? But when you create things that only you — with your unique talents and experience — can do, then you are absolutely not an imposter. You are the ordained. It is your duty to work on things that only you can do.
  10. Your best job will be one that you were unqualified for because it stretches you. In fact only apply to jobs you are unqualified for.
  11. It’s possible that a not-so smart person, who can communicate well, can do much better than a super smart person who can’t communicate well. That is good news because it is much easier to improve your communication skills than your intelligence.
  12. For the best results with your children, spend only half the money you think you should, but double the time with them.
  13. Don’t bother fighting the old; just build the new.
  14. You are as big as the things that make you angry.
  15. Efficiency is highly overrated; Goofing off is highly underrated. Regularly scheduled sabbaths, sabbaticals, vacations, breaks, aimless walks and time off are essential for top performance of any kind. The best work ethic requires a good rest ethic.

The observant among you might have noticed that I ended up picking fifteen rather than ten, but why short change myself and my readers? I didn’t bother culling out five – and to be clear, this is not to imply that the other eighty-eight are somehow inferior. These fifteen resonated the most with me, and I sincerely hope that your list is completely different from mine.

Note to self: of the ones I have selected here, the fifth one is the one where I really need to pull up my socks.

And speaking of hope, it would be nice if this list sparked conversations and your own lists!

Past mentions of Kevin Kelly on this blog are here.

David Warsh’s Take on Inflation

One of the sentences I have most enjoyed reading and internalizing is this one, by Scott Sumner: Never Reason From A Price Change.

I’ve capitalized each word in that sentence because it really is a sentence that makes you think until your head hurts. Here’s an early (perhaps the first) blog post from Scott in which he explains what he’s getting at:

My suggestion is that people should never reason from a price change, but always start one step earlier—what caused the price to change. If oil prices fall because Saudi Arabia increases production, then that is bullish news. If oil prices fall because of falling AD in Europe, that might be expansionary for the US. But if oil prices are falling because the euro crisis is increasing the demand for dollars and lowering AD worldwide; confirmed by falls in commodity prices, US equity prices, and TIPS spreads, then that is bearish news.

https://www.themoneyillusion.com/never-reason-from-a-price-change/

At its simplest – although there is always more to it than that – never reason from a price change means that the price might have changed because of demand, or supply or both. The headaches begin when you try to think through which of these might be more dominant, and the headache acquires splitting migraine status when you realize that you need to also ask about what else might be at play.

If you are a student of macroeconomics, a useful way to spend a morning is by clicking through this set of links and reading other posts by Scott Sumner on this topic. Remember, as always, the point is not to necessarily agree with Scott, but to read and ask how and why he arrives at his conclusions, and if you disagree with him, why do you do so. Best way to learn, especially if you can find a friend nerdy enough to do the exercise with you.


Which is a nice way to segue into our topic du jour: inflation.

A candy bar that cost a nickel in 1950 today costs $1.25 or so, depending on where you buy it. That, in a paper wrapper, is the price revolution of the twentieth century. Why did it happen? The answer usually given is that the quantity of money increased – too much paper money chasing too few candy bars.
A more satisfying explanation, casual though it may be, is to recognize that the global economy has grown considerably more complex since 1950, and the system of money, banking, and credit more complex along with it. The price of the candy bar wasn’t going to return to its previous level, no matter what the Fed or the candy-manufacturers did.

http://www.economicprincipals.com/issues/2022.05.01/2521.html

So begins a lovely little ruminative essay by David Warsh on how to think about inflation. It is lovely, but the emphasis in the previous sentence should be on the word “little”. I wish it was ten times longer!

But students used to textbook definitions of inflation might have their curiosity piqued after reading the second paragraph from the extract: what might complexity have to do with inflation?

A somewhat cryptic answer is given in the very next line that follows the end of the extract, where David Warsh refers to a book he wrote in 1984, called The Idea of Economic Complexity. I haven’t read the book, but I remember being told about it – alas, I can no longer remember who recommended it to me! But the idea of the book, from what I can recollect of the discussion, is as follows:

If you were to manufacture a Nokia 3310 today, odds are that you would be able to manufacture it at a fraction of the price that it commanded when it was first launched. Duh, you might think: so far, so obvious. Warsh’s point in the book is that this doesn’t necessarily mean that phones have become cheaper. In fact, as we can all attest, they go up in terms of price every year. The exact same thing might become cheaper, sure, but we keep making stuff more complex as we go along, and it is this increasing complexity that adds to inflation.

Now, bear in mind that I am treading on extremely thin ice over here! I’m describing a book to you that I haven’t read (strike one), on the basis of a conversation about the book that took place many years ago (strike two), and I’m now about to speculate on what else might be at play where this idea is concerned (strike three!).

All those CYA disclaimers aside, I’d like to think that complexity need not be just about the product itself, but could also be about the way it is manufactured, where all it is manufactured, where it is assembled, and how it is sold. Not to mention how all of this is financed!

As I’ve said before on these pages, macro is hard!


In Economic Development and the Price Level, in 1962, Geoffrey Maynard argued the opposite: that money generally adjusts to trade, rather than trade to money. In very different formats, the argument continues today.
“Development” is a bland word with which to describe the difference between the world economy in the time of Columbus and the world today. Economic philosopher David Ellerman has suggested that diversity describes the key difference, grounding his description in information theory; I proposed complexity in that 1984 book. But what is it that has become more diverse or complex? Not until I read “Increasing Returns an Economic Progress” (1928), by Allyn Young, did it occur to me that the growing complexity I had been thinking about were increases, of one sort or another, in the division of labor.

http://www.economicprincipals.com/issues/2022.05.01/2521.html

What a lovely excerpt, no? So much to add to the “To Read” list, but also how wonderful to pause and ponder on what the link might be between a Smithian division of labor and inflation. I hope you pause and think about this, much as I did when I read Warsh’s post, and again while drafting this paragraph right now.

To be clear: you’d expect division of labor to make systems more efficient, and therefore things cheaper. But Warsh suggests that there might be a way to link division of labor to complexity, and complexity to inflation!

Warsh ends his post in enigmatic fashion:

Are you comfortable with the too-much-money-chasing-too-few-goods story? Do you believe that the Fed could have prevented the rise in its price? And if wasn’t “inflation,” then what was it? The depreciation of money, relative to goods?
As with the sixteenth-century voyages of discovery, money follows development and development follows money. If you have only the quantity theory of money to rely on, you don’t know what is going on.

http://www.economicprincipals.com/issues/2022.05.01/2521.html

And that, I’d argue, is A Good Thing. A Good Thing because it allows us to prioritize reading The Idea of Economic Complexity, and allows to think about what David Warsh might be hinting at. An incentive (a carrot) to read the book, in other words, and one that I plan to use in the coming weeks.

AI/ML: Some Thoughts

This is a true story, but I’ll (of course) anonymize the name of the educational institute and the student concerned:

One of the semester end examinations conducted during the pandemic at an educational institute had an error. Students asked about the error, and since the professor who had designed the paper was not available, another professor was asked what could be done. Said professor copied the text of the question and searched for it online, in the hope that the question (or a variant thereof) had been sourced online.

Alas, that didn’t work, but a related discovery was made. A student writing that same question paper had copied the question, and put it up for folks online to solve. It hadn’t been solved yet, but the fact that all of this could happen so quickly was mind-boggling.

The kicker? The student in question had not bothered to remain anonymous. Their name had been appended with the question.

Welcome to learning and examinations in the time of Coviid-19.


I have often joked in my classes in this past decade that it is only a matter of time before professors outsource the design of the question paper to freelance websites online – and students outsource the writing of the submission online. And who knows, it may end up being the same freelancer doing both of these “projects”.

All of which is a very roundabout way to get to thinking about Elicit, videos about which I had put up yesterday.

But let’s begin at the beginning: what is Elicit?

Elicit is a GPT-3 powered research assistant. Elicit helps you classify datasets, brainstorm research questions, and search through publications.

https://www.google.com/search?q=what+is+elicit.org

Which of course begs a follow-up question: what is GPT-3? And if you haven’t discovered GPT-3 yet, well, buckle up for the ride:

GPT-3 belongs to a category of deep learning known as a large language model, a complex neural net that has been trained on a titanic data set of text: in GPT-3’s case, roughly 700 gigabytes of data drawn from across the web, including Wikipedia, supplemented with a large collection of text from digitized books. GPT-3 is the most celebrated of the large language models, and the most publicly available, but Google, Meta (formerly known as Facebook) and DeepMind have all developed their own L.L.M.s in recent years. Advances in computational power — and new mathematical techniques — have enabled L.L.M.s of GPT-3’s vintage to ingest far larger data sets than their predecessors, and employ much deeper layers of artificial neurons for their training.
Chances are you have already interacted with a large language model if you’ve ever used an application — like Gmail — that includes an autocomplete feature, gently prompting you with the word ‘‘attend’’ after you type the sentence ‘‘Sadly I won’t be able to….’’ But autocomplete is only the most rudimentary expression of what software like GPT-3 is capable of. It turns out that with enough training data and sufficiently deep neural nets, large language models can display remarkable skill if you ask them not just to fill in the missing word, but also to continue on writing whole paragraphs in the style of the initial prompt.

https://www.nytimes.com/2022/04/15/magazine/ai-language.html

It’s wild, there’s no other way to put it:


So, OK, cool tech. But cool tech without the ability to apply it is less than half of the story. So what might be some applications of GPT-3?

A few months after GPT-3 went online, the OpenAI team discovered that the neural net had developed surprisingly effective skills at writing computer software, even though the training data had not deliberately included examples of code. It turned out that the web is filled with countless pages that include examples of computer programming, accompanied by descriptions of what the code is designed to do; from those elemental clues, GPT-3 effectively taught itself how to program. (OpenAI refined those embryonic coding skills with more targeted training, and now offers an interface called Codex that generates structured code in a dozen programming languages in response to natural-language instructions.)

https://www.nytimes.com/2022/04/15/magazine/ai-language.html

For example:

(Before we proceed, assuming it is not behind a paywall, please read the entire article from the NYT.)


But about a week ago or so, I first heard about Elicit.org:

Watch the video, play around with the tool once you register (it’s free) and if you are at all involved with academia, reflect on how much has changed, and how much more is likely to change in the time to come.

But there are things to worry about, of course. An excellent place to begin is with this essay by Emily M. Blender, on Medium. It’s a great essay, and deserves to be read in full. Here’s one relevant extract:

There is a talk I’ve given a couple of times now (first at the University of Edinburgh in August 2021) titled “Meaning making with artificial interlocutors and risks of language technology”. I end that talk by reminding the audience to not be too impressed, and to remember:
Just because that text seems coherent doesn’t mean the model behind it has understood anything or is trustworthy
Just because that answer was correct doesn’t mean the next one will be
When a computer seems to “speak our language”, we’re actually the ones doing all of the work

https://medium.com/@emilymenonbender/on-nyt-magazine-on-ai-resist-the-urge-to-be-impressed-3d92fd9a0edd

I haven’t seen the talk at the University of Edinburgh referred to in the extract, but it’s on my to-watch list. Here is the link, if you’re interested.

And here’s a Twitter thread by Emily M. Blender about Elicit.org specifically:


In response to this critique and other feedback, Elicit.org have come up with an explainer of sorts about how to use Elicit.org responsibly:

https://ought.org/updates/2022-04-25-responsibility

Before we proceed, I hope aficionados of statistics have noted the null hypothesis problem (which error would you rather avoid) in the last sentence of pt. 1 in that clipping above!


So all that being said, what do I think about GPT3 in general and elicit.org in particular?

I’m a sucker for trying out new things, especially from the world of tech. Innocent until proven guilty is a good maxim for approaching many things in life, and to me, so also with new tech. I’m gobsmacked to see tools like GPT3 and DallE2, and their applications to new tasks is amazing to see.

But that being said, there is a lot to think about, be wary of and guard against. I’m happy to keep an open mind and try these amazing technologies out, while keeping a close eye on what thoughtful critics have to say.

Which is exactly what I plan to do!

And for a person with a plan such as mine, what a time to be alive, no?

Have you tried Elicit.org yet?

Video 1:

And Video 2: