Month: March 2023

What is the single best way of improving your GPT prompts?

by  Tyler Cowen in

I have a nomination, and here is an excerpt from my new paper with Alex:

You often can get a better and more specific answer by asking for an answer in the voice of another person, a third party. Here goes: What are the causes of inflation, as it might be explained by Nobel Laureate Milton Friedman?

By mentioning Friedman you are directing the GPT to look at a more intelligent segment of the potential answer space and this directing will usually get you a better answer than if you just ask “What are the causes of inflation?” Similarly, you want all of the words used in your query to be intelligent-sounding. Of course, you may not agree with the views of Friedman on inflation. Here are a few economists who are well known and have written a lot on a wide variety of issues:

Paul Samuelson
Milton Friedman
Susan Athey
Paul Krugman
Tyler Cowen
Alex Tabarrok

But you don’t have to memorize that list, and it is not long enough anyway. When in doubt, ask GPT itself who might be the relevant experts. How about this?: “I have a question on international trade. Which economists in the last thirty years might be the smartest experts on such questions?” The model will be very happy to tell you, and then you can proceed with your further queries.

Of course this advice generalizes far beyond economics.  A friend of mine queried GPT-4 about Jon Fosse, a Norwegian author, and received a wrong answer.  He retried the same question, but asking also for an answer from a Fosse expert.  The response was then very good.

The title of my paper with Alex is “How to Learn and Teach Economics with Large Language Models, Including GPT,” but again most of the advice is generalizable to education with GPTs more generally.  Recommended, the paper is full of tips for using GPT models in more effective ways.

Imagine if humanity ends up divided into two classes of people: those who are willing/not embarrassed to tack on extra “silly bits” to their prompts, and those who are not so willing.  The differences in capabilities will end up being remarkable.  Are perhaps many elites and academics unwilling to go the extra mile in their prompts?  Do they feel a single sentence question ought to be enough?  Are they in any case constitutionally unused to providing extra context for their requests?

Time will tell.

Lots of announcements from Glen Weyl

by  Tyler Cowen in

Sunday assorted links

by  Tyler Cowen

1. Claims about Neom (speculative, three-minute video).

2. Keep your eye on the bank holding company.

3. This guy was described to me as “the Michael Orthofer of cinema.”

4. Explaining word embeddings in LLMs.

5. GPT-4’s multi-modal capabilities: a preview.

6. Daniel Gross/Nat Friedman podcast on using AI to decipher ancient scrolls.  With Brent Seales, recommended.

7. Orson Welles on innovation and creative control (short video).

8. Haitian art in Kassel.

Chat Law Goes Global

by  Alex Tabarrok in

PricewaterhouseCoopers (PWC), the global business services firm, has signed a deal with OpenAI for access to “Harvey”, OpenAI’s Chatbot for legal services.

Reuters: PricewaterhouseCoopers said Wednesday that it will give 4,000 of its legal professionals access to an artificial intelligence platform, becoming the latest firm to introduce generative AI technology for legal work.

PwC said it partnered with AI startup Harvey for an initial 12-month contract, which the accounting and consulting firm said will help lawyers with contract analysis, regulatory compliance work, due diligence and other legal advisory and consulting services.

PwC said it will also determine ways for tax professionals to use the technology.

IBM’s Watson was a failure so we will see but, yeah I will say it, this time feels different. For one, lawyers deal with text where GPTs excel. Second, GPTs have already revolutionized software coding and unlike Watson I am using GPTs every day for writing and researching and it works. The entire world of white collar work is going to be transformed over the next year. See also my paper with Tyler, How to Learn and Teach Economics with Large Language Models, Including GPT.

Where are all the Indonesians?

by  Tyler Cowen

Here is an old reader email query from 2015:

I started by asking, why are there so few Indonesians in the US? The email subject is just a juicy comparison. Indonesians are outnumbered in the U.S. by 14 other Asian-American sub-groups. Most estimates give an answer of about 100,000 total, this from a country of 250+ million.

Theories:

History – Indonesia’s colonial experience is Dutch (although less than 500,000 live in the Netherlands). There do not seem to be many self-identifying South African Americans. The country has more recent independence/consolidation than some.

Religion – Islam reduces immigration demand and supply

Economics – Indonesians are very poor and/or less skilled for particular types of migrant labor. Perhaps why so many are in the Middle East (maybe 1.5 million in Saudi Arabia, although this seems more a recent phenomenon).

Internal markets – Indonesia is large and diverse. Opportunity and adventure are an island away, not a country.

Conflict – Sukharno/Suharto rule uniquely dampened emigration.

Reporting – The range of Indonesian ethnicities is not suitable to census counts. Ask Sir Jervoise Baines about this.

Some combination of the Philippines, Pakistan, India, and Myanmar nix all of the above. The more I think and ask others, the more my answer is “the sum of the valid remainder of all the other explanations.” It is not satisfying.

I am in Indonesia now and ask the question often. Most Indonesians don’t know that there are so few, relatively speaking. One told me about the roughly 500 Indonesians at her alma mater Ohio State University, home of the Center of SouthEast Asia Studies and Professor R. William Liddle (that is why she went there).

Bailouts of insolvent banks don’t lead to hyperinflation

by  Tyler Cowen

Let’s say there is a big hole in the solvency of a banking system.  Left unaddressed, that is radically deflationary.  Demand (and other) deposits will disappear, crushing aggregate demand.  Cascading financial failures will occur elsewhere, again with negative demand effects.  If a government “prints money,” or more likely creates new electronic bookkeeping entries, that offsets the deflationary pressures.  These bailouts may have other negative effects, such as on future moral hazard and rent-seeking, but they won’t bring hyperinflation.  If you wanted to create hyperinflation, the bailout would have to look something like “for every dollar you used to have in your bank account, the Fed says you now have five!”  But that is not on the agenda.

Saturday assorted links

by  Tyler Cowen

1. Who should next run the NIH?

2. Did 18th century British firefighters just let buildings burn if they hadn’t paid up?  Short video.

3. “Developers demolished a historic pub. They must rebuild from the rubble.”  The heritage culture that is England.  I recommend the photo of the pile of rubble.  Was the original building so attractive in the first place?

4. “The Microsoft Copilot “future of work” demo is incredible. Your boss will soon be able to ask their Copilot to create a summary of who does the least work on average and have their termination letter already drafted in Outlook.”  Here is the Sam Hammond tweet.

5. Ethan Mollick paper on effective teaching strategies with GPT.  And GPT-4 is playing chess at an OK level, though it was never taught that skill.

6. Is Microsoft adding a crypto wallet to its Edge browser?

Teaching and Learning Economics with the AIs

by  Alex Tabarrok in

Tyler and I have a new paper, How to Learn and Teach Economics with Large Language Models, Including GPT:

GPTs, such as ChatGPT and Bing Chat, are capable of answering economics questions, solving specific economic models, creating exams, assisting with research, generating ideas, and enhancing writing, among other tasks. This paper highlights how these innovative tools differ from prior software and necessitate novel methods of interaction. By providing examples, tips, and guidance, we aim to optimize the use of GPTs and LLMs for learning and teaching economics effectively.

Most of the paper is about how to use GPTs effectively but we also make some substantive points that many people are missing:

GPTs are not simply a chatty front end to the internet. Some GPTs like ChatGPT have no ability to search the internet. Others, like Bing Chat, can search the internet and might do so to aid in answering a question, but that is not fundamentally how they work. It is possible to ask a GPT questions that no one has ever asked before. For example, we asked how Fred Flintstone was like Hamlet, and ChatGPT responded (in part):

Fred Flintstone and Hamlet are two vastly different characters from different time periods, cultures, and mediums of storytelling. It is difficult to draw direct comparisons between the two.

However, one possible point of similarity is that both characters face existential dilemmas and struggles with their sense of purpose and identity. Hamlet is plagued by doubts about his ability to avenge his father’s murder, and his own worthiness as a human being. Similarly, Fred Flintstone often grapples with his place in society and his ability to provide for his family and live up to his own expectations.

Not a bad answer for a silly question and one that (as far as we can tell) cannot be found on the internet.

GPTs have “read” or “absorbed” a great amount of text but that text isn’t stored in a database; instead the text was used to weight the billions of parameters in the neural net. It is thus possible to run a GPT on a powerful home computer. It would be very slow, since computing each word requires billions of calculations, but unlike storing the internet on your home computer, it is feasible to run a GPT on a home computer or even (fairly soon) on a mobile device.

GPTs work by predicting the next word in a sequence. If you hear the phrase “the Star-Spangled”, for example, you and a GPT might predict that the word “Banner” is likely to come next. This is what GPTs are doing but it would be a mistake to conclude that GPTs are simply “autocompletes” or even autocompletes on steroids.

Autocompletes are primarily statistical guesses based on previously asked questions. GPTs in contrast have some understanding (recall the as if modifier) of the meaning of words. Thus GPTs understand that Red, Green, and Blue are related concepts that King, Queen, Man and Woman are related in a specific way such that a woman cannot be a King. It also understands that fast and slow are related concepts, such that a car cannot be going fast and slow at the same time but can be fast and red and so forth. Thus GPTs are able to “autocomplete” sentences which have never been written before, as we described earlier.2 More generally, it seems likely that GPTs are building internal models to help them predict the next word in a sentence (e.g. Li et al. 2023).

The paper is a work in progress so comments are welcome.

What do I think of the economists’ Israel petition?

by  Tyler Cowen in

It is signed by many luminaries, and it opens with this:

The governing coalition in Israel is considering an array of legislative acts that would weaken the independence of the judiciary and its power to constrain governmental actions. Numerous Israeli economists, in an open letter that some of us joined, expressed concerns that such a reform would adversely affect the Israeli economy by weakening the rule of law and thereby moving Israel in the direction of Hungary and Poland. Although we significantly vary in our views on public policy and on the challenges facing Israeli society, we all share these concerns. A strong and independent judiciary is a critical part of a system of checks and balances. Undermining it would be detrimental not only to democracy but also to economic prosperity and growth.

I would say I haven’t made up my mind on the substantive issue, as I have seen credible (not saying they are true, I don’t know) arguments that the current Israeli judiciary has too much power.  The proposed reforms still might be a badly timed and significant overreach, but my intuition is that the arguments are more complicated than this petition is making them out to be.  As economists, are they not at least obliged to tell us what the relevant trade-off is?

I also wonder if these outside voices have influence in Israeli politics, or whether they might occasion backlash.  Again, I don’t know, but I do see an argument for reserving collective petitions for very clear cut cases when the transmitted signal will be positive.  Is the binding constraint here “not having enough elite academic foreigners in opposition to Netanyahu”?

More generally and perhaps most importantly, will this petition be effective?  Many kinds of petitions should be saved up for when they will change something.  If they are not going to matter, in essence the signers are signaling their weakness rather than their strength.  They are spending down their reputational capital, rather than building it up.  And in those cases, why have the petition at all?

Steven Pinker on existential risk

by  Tyler Cowen in

He is harsh, but my view is not far from his:

The AI-existential-threat discussions are unmoored from evolutionary biology, cognitive psychology, real AI, sociology, the history of technology and other sources of knowledge outside the theater of the imagination. I think this points to a meta-problem. The AI-ET community shares a bad epistemic habit (not to mention membership) with parts of the Rationality and EA communities, at least since they jumped the shark from preventing malaria in the developing world to seeding the galaxy with supercomputers hosting trillions of consciousnesses from uploaded connectomes. They start with a couple of assumptions, and lay out a chain of abstract reasoning, throwing in one dubious assumption after another, till they end up way beyond the land of experience or plausibility. The whole deduction exponentiates our ignorance with each link in the chain of hypotheticals, and depends on blowing off the countless messy and unanticipatable nuisances of the human and physical world. It’s an occupational hazard of belonging to a “community” that distinguishes itself by raw brainpower. OK, enough for today – hope you find some of it interesting.

That is by no means the only harsh paragraph.  Here is the entire dialogue with Richard Hanania.  And be careful what you write in the MR comments section, the AIs are reading you!

What should I ask Jonathan Swift?

by  Tyler Cowen in

Yes, I would like to do a Conversation with Jonathan “G.P.T.” Swift.  Here is Wikipedia on Swift, excerpt:

Jonathan Swift (30 November 1667 – 19 October 1745) was an Anglo-Irish satirist, author, essayist, political pamphleteer (first for the Whigs, then for the Tories), poet, and Anglican cleric who became Dean of St Patrick’s Cathedral, Dublin, hence his common sobriquet, “Dean Swift”.

Swift is remembered for works such as A Tale of a Tub (1704), An Argument Against Abolishing Christianity (1712), Gulliver’s Travels (1726), and A Modest Proposal (1729). He is regarded by the Encyclopædia Britannica as the foremost prose satirist in the English language.[1] He originally published all of his works under pseudonyms—such as Lemuel Gulliver, Isaac Bickerstaff, M. B. Drapier—or anonymously. He was a master of two styles of satire, the Horatian and Juvenalian styles.

His deadpan, ironic writing style, particularly in A Modest Proposal, has led to such satire being subsequently termed “Swiftian”.

So what should I ask him?  I thank you in advance for your suggestions.

Singaporean hawker centre in Manhattan

by  Tyler Cowen in

Urban Hawker, On 135 W.50th, 17 vendors  Here is a NYT review, good photos of the key dishes.  The Hainanese chicken rice was amazing, worthy of Singapore, get it poached of course.  Condiments!  The Malaysian lontong was quite good, the beef rendang decent.  The lamb biryani I enjoyed, with a thick sauce than you would not find in Hyderabad, laden with cloves and cinnamon.  Most of the people there are not Singaporean, but many have “that Singaporean look,” so it feels fairly authentic, except for the prices, which run about $20 a course.  Ordering your meal and finding/keeping a table can be difficult, also making it authentic.  (Choping needed!)  Ordering a meal and getting a drink of water on the same trip can be difficult, making it more authentic yet.  Overall, not as good as it could be but better than you might be expecting.  Some of the vendors verge on Pan-Asian rather than Singaporean proper, but ultimately Singapore itself is headed in that direction.  So I will go again, though I can’t imagine the chili crab is worth the price.  Most of all, you need to go early rather than at peak times.

And if you are wondering what “that Singaporean look” means, I suppose it refers to looking down a bit, earnest, and seeming not entirely happy, all the while focused on getting some excellent food.