Diplomacy is a 7-player game in which players must persuade, cajole, coordinate, strategize, bluff and lie to one another in order to take over the world. For the first time, an AI has achieved success in Diplomacy:
Over 40 Diplomacy games with 82 human players involving 5,277 messages over 72 hours of gameplay, CICERO achieved more than double the average score of the other players and ranked in the top 10% of players!
Note that this AI isn’t just a large language model, it’s a strategic engine connected to a language model–thus it figures out what it wants to do and then it convinces others, including gaining sympathy, bluffing and lying, to get others to do what it wants to do.
Here’s some correspondence from one game. Can you tell which is the AI?
CaptainMeme, a professional Diplomacy player, runs through an entire blitz game here. What’s interesting is that he hardly comments on the AI aspect and just treats it as a game with 6 other very good players.
Paper and more discussion here. Keep in mind that since the game is zero-sum to do well the AI must convince humans to do what is NOT in their interest. We really do need to invest more in the alignment problem.
We introduce a model of the admissions process based upon standard agency theory and explore its implications with economics PhD admissions data from 2013-2019. We show that a subjective score that aggregates subjective ratings and recommendation letter features plays a more important role in determining admissions than an objective score based upon graduate record exam (GRE) scores. Subjective evaluations by references who write multiple letters are not only more influential than those of references who write one letter, but they are also more informative. Since multiple-letter references are also more highly ranked economists, this implies that there is a constraint on the supply of high-quality references. Moreover, we find that both the subjective and objective scores are correlated with job placement at a top economics department after the completion of the PhD. These indicators of individual achievement have a smaller effect than an undergraduate degree from an Ivy Plus school (i.e., Ivy League + Stanford, MIT, Duke, and Chicago). In the self-selected pool of applicants, Ivy Plus graduates are twice as likely to be admitted to a top 10 graduate program and are much more likely to obtain an assistant professor position at a top 10 program upon PhD completion. Given that Ivy Plus students must pass a stringent selection process to gain admission to their undergraduate program, we cannot reject the hypothesis that admission committees use information efficiently and fairly. However, this also implies that there may be a return to attending a selective undergraduate program in order to be pooled with highly skilled individuals.
Here’s my high-level explanation of the FTX crash.
Imagine that I own a house and I create a million coins representing the value of the house. I give half of the coins to my wife. I then sell one of my coins to my wife for $10. Now the house has a nominal value of $10 million dollars and my wife and I each have assets worth $5 million. Of course, no one is likely to buy my house for $10 million or lend me money based on my coin wealth but suppose I now get my friend Tyler to buy a coin for $15. Tyler says why would I want to buy your s!@# coin! To encourage Tyler to buy I give him a side-deal that is not very public. Say an extra 5% of our textbook royalties. Tyler buys the coin for $15. Now the coins have gone up in value by 50%. My wife and I each have $7.5 million. Other people may want to get in while they can—Tyler bought in! Are you in? I’m in!
Now if it’s not obvious, I am SBF in the analogy, and my wife is Alameda run by his sometimes girlfriend Caroline Ellison. Who is Tyler?—the seeming outsider who gets a kind of under-the-table deal to pump SBF’s coins? One possibility, is Sequoia a venture capitalist firm who invested in FTX, SBF’s house, while at the same time FTX invested in Sequoia. Weird right? Tyler in this example is also a bunch of firms that Alameda invested in but which were then required to keep their funds at FTX. Many other possibilities exist.
Another relevant point to our analogy is that there are one million coins but only a handful of them are traded, the handful that are traded are called the float. Similarly, many crypto coins were created with emissions schedules where only a few coins were released, the float, with a majority of the coins “locked” and only released over time. Keeping the price high, and thus the imputed value of the stock high, meant you only had to control the float.
Ok, so far this is crazy but despite nominal values in the millions a relatively small amount of real money has actually changed hands. But suppose that I now open a bank or an exchange. People want to bank with me since I have clearly shown that I know how to get wealthy! Now the money coming into the exchange is real money and it’s a bull market so when people check their accounts everything looks great, everyone is making money.
Suppose I take some of these assets and lend them to my wife for her to take speculative bets on. Is this illegal? Well, it’s actually hard to say. A bank is supposed to make loans. It’s more complicated with an exchange. Maybe it’s illegal, maybe not. After all when I lend assets to my wife I can say that there was lots of collateral. What collateral? Well remember my wife has $7.5 million in coins so I am lending say $3 or $4 million which is backed by twice as much collateral—that looks safe, right? Actually, it’s even better since she is going to invest the assets in other assets, unfortunately other coins not the S&P500, but now there is even more collateral. Everything looks safe.
Importantly, if the assets my wife is investing in are going up in price—she is getting very, very rich. She borrowed billions and keeps all the profits on the upside. Give me a house of assets to stand on and with leverage I will rule the earth! Moreover, the more prices go up, the safer this trade looks since the collateral is increasing in value. Also, my wife and I can coordinate on which coins to buy. She buys and then I list the coins on my exchange and offer them to all my customers. More demand, more price appreciation, more demand. My wife decides to borrow even more, since the trade is working so well.
Ok, now we get to the end of 2021 and what happens? After a massive run up in prices, crypto price start dropping.* Other firms in the space including Voyager and BlockFi start to come under pressure because of the TerraUSD-Luna collapse in May of 2022. Now, the bets aren’t starting to look so good. So what do I do. Either I come clean and reorganize or double down. It looks like SBF doubled down. More borrowing and more big bets. Amazingly, SBF offered to buy Voyager and BlockFi and bail them out. At the time, this looked like a visionary move to save crypto. Finance experts compared SBF to JP Morgan, the private banker who took big bets in 1907 to reestablish confidence like a proto-central bank. What we learned later, however, was that SBF owed these firms money and if they started to demand payment that would put pressure on his collateral, the coins on the house that we talked about earlier. So SBFs efforts to buy these firms were an effort to keep his own weakness hidden. Indeed, as people start to sell their coins, Alameda had to step in to buy, to keep the price up.
Eventually, as people began to look more closely at the assets of Alameda and FTX they realized that many of the numbers were huge stock-valuations made on tiny floats–not just the original house coins but also many of the coins, like Serum, bought by Alameda as investments. And once people realized that, they ran to get out before the house burned down. Now everything works in reverse—a $10 trade goes to $1 and your valuation is cut by billions overnight. We also get fire sales—as firms try to sell assets to meet their customer demands the prices of those assets fall which makes people sell other assets and so the contagion spreads (as described in Modern Principles).
Ok, final analogy. Suppose to help me run my house I invite over a bunch of friends and we do a lot of drugs and hook up together and suppose that none of us really knows anything about accounting or financial controls.
Well that about covers it.
N.B. Much of this story is familiar. The assets involved were crypto tokens but they could have been fiat currencies, internet stocks or mortgage backed securities. The new and original aspects of cryptofinance such as decentralized consensus, crypto wallets, and automated marker makers continue to work well. Unfortunately, these fine distinctions are not likely to be widely understood.
—
*You might wonder why crypto prices started dropping. One important reason is macroeconomic, rising interest rates. When interest rates are very low a dollar in the far future is worth almost as much as a dollar today. Thus, in a regime of low interest-rates, crypto and other projects with (speculative) long-run payoffs could be valued highly. As interest rates rose, however, long-run speculative returns began to look much less attractive than say T-bills and money flocked out of assets in the long-run sector causing prices to plummet.
An interesting piece in The Guardian by Howard French on Africa’s megalopolis and the difficulties of pulling together five countries with very different governments and colonial histories:
There is one place above all that should be seen as the centre of this urban transformation. It is a stretch of coastal west Africa that begins in the west with Abidjan, the economic capital of Ivory Coast, and extends 600 miles east – passing through the countries of Ghana, Togo and Benin – before finally arriving at Lagos. Recently, this has come to be seen by many experts as the world’s most rapidly urbanising region, a “megalopolis” in the making – that is, a large and densely clustered group of metropolitan centres.
…In just over a decade from now, its major cities will contain 40 million people. Abidjan, with 8.3 million people, will be almost as large as New York City is today. The story of the region’s small cities is equally dramatic. They are either becoming major urban centres in their own right, or – as with places like Oyo in Nigeria, Takoradi in Ghana, and Bingerville in Ivory Coast – they are gradually being absorbed by bigger cities. Meanwhile, newborn cities are popping into existence in settings that were all but barren a generation ago. When one includes these sorts of places, the projected population for this coastal zone will reach 51 million people by 2035, roughly as many people as the north-eastern corridor of the US counted when it first came to be considered a megalopolis.
But unlike that American super-region, whose population long ago plateaued, this part of west Africa will keep growing. By 2100, the Lagos-Abidjan stretch is projected to be the largest zone of continuous, dense habitation on earth, with something in the order of half a billion people.
To the extent crypto clearinghouses and exchanges have a future, they too will be regulated, and this is all the more certain after the FTX fiasco. Then the question becomes: How many of the (supposed) efficiencies of crypto would remain under such a regulated regime? After all, the original point of crypto was to lower the transaction costs associated with traditional financial institutions. Intermediary costs, reserve requirements and legal compliance costs could more than reverse those advantages.
Intermediaries nonetheless have proliferated in crypto, for some obvious reasons. Quite simply, most people do not want to have to deal with the trouble of running their own crypto wallet, safeguarding their password and figuring out how the system works. It is daunting, even for people sophisticated about finance or technology.
Now enter AI. New AI systems are getting very good at voice recognition, at executing commands, at understanding text, and even at writing their own computer programs. Is it such a stretch to imagine an AI that makes a crypto wallet easy to use?
You would still hold your crypto in your own wallet, and would not need to trust any intermediary, except of course for the AI itself. At will, you would give your AI desired commands. Open a wallet for me. Send 0.1 Bitcoin to my brother. Convert all my accounts into cash. And so on.
In essence, the AI would ease your interactions with the system, but without creating a separate corporate entity between you and your funds. If the AI company went bankrupt, your funds would still be in your wallet. Probably the AI program would manage your personal finances more broadly, not just your crypto wallet.
You might wonder whether you could trust the company supplying the AI. But that question is answered relatively easily with another: Do you trust your smartphone or computer to do online banking? For the vast majority of people, the answer is yes. But if those companies built software programs to intercept or redirect consumer funds flows for their own purposes, those attempts would not last a day and the companies would rapidly be out of business and in court.
There are some obvious specific causes behind the FTX debacle, but it also reflects some more general problems with the clearinghouse/exchange business model.
are very good reviews and summaries of where we stand. The third discusses what Brink calls The Anti-Promethean Backlash
…the anti-Promethean backlash — the broad-based cultural turn away from those forms of technological progress that extend and amplify human mastery over the physical world. The quest to build bigger, go farther and faster and higher, and harness ever greater sources of power was, if not abandoned, then greatly deprioritized in the United States and other rich democracies starting in the 1960s and 70s. We made it to the moon, and then stopped going. We pioneered commercial supersonic air travel, and then discontinued it. We developed nuclear power, and then stopped building new plants. There is really no precedent for this kind of abdication of powers in Western modernity; one historical parallel that comes to mind is the Ming dynasty’s abandonment of its expeditionary treasure fleet after the voyages of Zheng He.
This is a chart of U.S. energy consumption per capita, which until around 1970 showed steady exponential growth of around 2 percent a year. The author calls this the “Henry Adams curve,” since the historian was an early observer of this phenomenon. But around 1970, the Henry Adams curve met the anti-Promethean backlash — and the backlash won.
The chart comes from Where Is My Flying Car? by J. Storrs Hall. It’s a weird, wacky book that rambles all over the place; it’s also brilliant, and it changed my mind about a matter of great importance.
Before reading Hall, if I had seen this chart — and maybe I did see something like it before, I’m not sure — I would have had a completely different reaction. My response would have been along the lines of: “Wow, look at capitalism’s ever-increasing energy efficiency. We’re getting more GDP per kilowatt-hour than ever before, thanks to information technology and the steady dematerialization of economic life. All hail postmaterialist capitalism!”
But Hall argues convincingly that the plateauing of the Henry Adams curve didn’t represent the natural evolution of capitalism in the Information Age. The bending of that curve, he claims, constituted self-inflicted injury. Our midcentury dreams of future progress — flying cars, nuclear power too cheap to meter, moon bases and underwater cities — didn’t fail to materialize simply because we were lousy at guessing how technology would actually develop. They failed to materialize because the anti-Promethean backlash, aided by with loss-averse apathy, left them strangled in their cribs.
He is especially convincing on nuclear power. My prior impression was that nuclear power had always been a high-cost white elephant propped up only by subsidies, but Hall documents that back in the 1950s and 60s, the cost of new plants was falling about 25 percent for every doubling of total capacity — a classic learning-curve trajectory that was abruptly halted in the 1970s by suffocating regulation. In 1974 the Atomic Energy Commission was abolished and the new Nuclear Regulatory Commission was established. In the almost half-century since then, there has not been a single new nuclear power plant approved and then subsequently built.
How should one incentivize creativity when being creative is costly? We analyze a model of delegated bandit experimentation where the principal desires the agent to constantly switch to new arms to maximize the chance of success. The agent faces a fixed cost of switching. We show that the principal’s optimal reward scheme is maximally uncertain—the agent receives transfers for success, but their distribution has an extreme variance. Despite being stationary, the optimal reward scheme achieves the principal’s first-best outcome provided that the agent’s outside option is sufficiently valuable. Our results shed light on the non-transparent incentives used by online platforms, such as YouTube, and guide how to design incentives for creativity in such applications.
One feature of this model is that extreme uncertainty about rewards motivates project-switching, which is what the principle wants. Most projects should offer low rewards, but a small percentage of winners should offer very high rewards. In this model it is also the case that opaque bonus schemes perform better than transparent ones. As I understand this result, the principal wants the agent to keep on switching and thus does not want to offer any kind of “safe haven” where the agent can rest securely.
The upshot is that there is a tendency for members of a clearinghouse to either a) fail to meet standards and go bust, or b) join or least collude with a dominant coalition.
And:
You could also argue that a dominant clearinghouse might be good for crypto. The history of banking includes dominant or semi-dominant clearinghouses stabilizing markets and helping to introduce innovations, for instance of timeliness and transparency. The collusive monopoly might take too big a share of the market surplus for itself, but it has an incentive to keep the market up and running and profitable. That is hardly the worst arrangement crypto might stumble upon.
It is also true that a dominant clearinghouse is much easier to regulate, and indeed modern central banks often sprung out of these earlier clearinghouse arrangements. Sooner or later, there is a tendency for the law to intervene and turn the dominant private clearinghouse into part of a more formalized central bank. It is no accident that member banks of the Fed are still called “stockholders.”
One complication is that Binance is not a US firm; incorporated in China, it is now based in Dubai. Regulators might hope an American or at least Western version of Binance comes along, perhaps to create a new market duopoly. Arguably that is what regulators were hoping all along for FTX, so at least one version of the previous plan now has a huge hole in it. All the more pressure will be placed on Coinbase (a US firm), which may gain business but face a heavier regulatory burden and be expected to play a more specific role in the system.
When guessing at the future of crypto, keep in mind that the future of crypto exchanges and the future of crypto assets are very different things. For many pure crypto bugs, the exchanges are a sellout and a concession to older methods of finance and settlement. The exchanges can be regulated, controlled and co-opted, even turned against the notion of individual monetary sovereignty. Instead, the pure crypto vision stresses the notion of “every person their own bank,” through the medium of a personal wallet and beyond easy purview of the central authorities.
When confronted with a puzzle, economists instintively look for an explanation based on the incentives of rational people acting within institutional constraints. The economic approach is humble–rather than assuming that something is wrong with other people, the economists begins to unravel a puzzle by assuming that there is something the economist doesn’t understand. The economic approach is respectful of other people’s preferences and the constraints that they operate under. The economic approach is egalitarian, it looks for an answer that amounts to, “ah, now that I understand this, if I were in the same position I would do as these people are doing.” The economic approach is powerful but it can be taken too far. Sometimes people aren’t rational. As Larry Summer said, “There are idiots. Look around.”
In Folk Economics and the Persistence of Political Opposition to New HousingClayton Nall, Chris Elmendorf, and Stan Oklobdzija look around at the housing market and declare there are idiots. Well, they are more polite but that’s the gist. Did we really need a 96 page paper to tell us this? Well, I didn’t. I was convinced of the problem years ago when after giving a rational explanation of why it’s difficult to get new housing approved, Bryan Caplan remarked wryly “yeah, that’s all true but renters are also opposed to new housing!” What this paper shows is that Bryan was right and that it matters.
Using two nationally representative surveys of urban and suburban residents, with embedded experiments, we show that about 30%-40% of Americans believe, contrary to basic economic theory and robust empirical evidence, that a large, exogenous increase in their region’s housing stock would cause rents and home prices to rise. This finding is robust to variations in the source of the supply shock, the manner in which price predictions are elicited, and the stated assumptions about counterfactual future prices in the absence of the shock. Moreover, by comparing homeowners and renters, by relating price predictions to self-reported confidence in the prediction, by comparing the predicted effects of more- and less-popular supply-shock scenarios, and by exploiting a survey experiment, we are able to discount motivated reasoning or guessing as explanations for these responses.
The authors do have new results which I wouldn’t have predicted. People are idiots about housing especially but not about rutabagas.
Housing “Supply Skepticism” is not just a manifestation of general economic ignorance. We show that the public understands the implications of supply and demand in markets for agricultural commodities, for labor, and even for cars, a durable consumer good that, like housing, trades in new and second-hand markets. There is also overwhelming agreement about how home prices and rents are locally affected by changes in neighborhood quality, by in-migration of rich people, by expansions of employment, by demolition of affordable homes, by new construction of expensive housing next door to more affordable homes, and (perhaps more questionably) by corporate ownership.
A deep problem is that people tend to blame developers of housing for high prices, i.e. they blame high prices on the one group most responsible for lowering prices! We see the same thing in other markets, of course, but in housing this effect appears especially strong.
…we observe a very strong tendency to blame housing providers (developers) for high housing prices. Conversely, actors whose stock in trade is opposing new development (environmentalists, anti-development activists) are almost never blamed. Incumbent owners who rent out their property (landlords) are commonly blamed, whereas incumbent owners who occupy it (homeowners) are not, even though both groups have a similar economic interest in supply restrictions. The pattern of blame attribution is very similar for homeowners and renters, except that renters hold landlords more responsible and homeowners are slightly more likely to list developers in their top three.
Addendum: Lots of folks are commenting in the comments.
One of the best biographies of any economist, covers anything you might wish to know, and with conceptual understanding. This is a fantastic book and I am eagerly awaiting volume II. Mini-excerpt:
Hayek [at LSE] never really liked Mannheim, whom he spent some time with when he first arrived, trying to introduce him around and help him to get acclimated. This ended when, after listening patiently to Mannheim complain about the inadequacy of the English language for expressing his ideas, Hayek finally blurted out, “So much the worse for your ideas!”
You can pre-order here. Hayek, by the way, is an interesting polar case for any talent search algorithm. He was first interested in botany, and didn’t do anything in economics until he was 30 years old.
And I hadn’t known that Colette was the stepmother of Bertrand de Jouvenal (“On Power”), and had a five-year affair with him!
Say the price of oil is $100 per barrel, and extraction costs Russia $50 per barrel. The profit on that oil is $50 per barrel. Now assume a price cap of $70 per barrel. Russian profit falls to $20 per barrel — but still, the oil will be produced. A finely tuned price control would redistribute income toward buyers of Russian oil, without much interfering with oil supply.
Enforcement would have to be through maritime services, such as insurance for the carrying ships.
But I am not sold on its workability or efficacy:
One problem is how to set the cap at the right level. The plan is to set a fixed cap, rather than at some percentage discount to world oil prices. As world oil prices change, it would therefore be necessary to adjust the cap, preferably quickly. Given that it has taken months to agree to the idea of a cap, it remains to be seen if this would be possible. And things only get worse if the Western coalition against Russia splinters, or if the relevant bureaucracies are slow.
If the price cap ends up too low and Russian oil is taken off world markets altogether, that could significantly worsen what is already a serious global economic downturn.
A second problem is that Russia might simply sell the oil to nations not participating in this agreement to cap prices, most notably China and India. But selling more to those countries might require Russia to lower the price. And while China and India are unlikely to join the G-7 plan, the very existence of the price cap gives them bargaining leverage over Russia.
The bottom line, however, remains: Any decline in Russian government revenue might be considerably less than what a plan to cap the price of oil might indicate. And that’s not even considering whatever Russia might earn from selling oil on the black market. Nations outside the G-7 would have an incentive to buy tankers, self-insure them and use them to ship Russian oil without the price cap.
But the major issue is one of escalation…
In any case worth a ponder, there are further arguments in the piece.
I think I forgot to mention this when they won the Nobel Prize. In their model of bank runs, there are multiple equilibria and people can run on the bank simply if they expect other depositors will run on the bank too. The combination of liquid liabilities and illiquid assets then means the bank cannot meet all their claims. That is logically consistent, but I think not realistic. Virtually all the bank runs I know of stem from insolvency, not rumors, noise, and multiple equilibria. Didn’t Hugh Rockoff do some papers showing that the “free banking era” bank runs were pretty rational in the sense that the depositors targeted the right institutions? Or did someone other than Rockoff do this work? In any case, I’ve long thought that bank run models based on insolvency are more useful than bank run models based on rumors and multiple equilibria.
Switzerland, one of the world’s richest nations, has an ambitious climate goal: It promises to cut its greenhouse gas emissions in half by 2030.
But the Swiss don’t intend to reduce emissions by that much within their own borders. Instead, the European country is dipping into its sizable coffers to pay poorer nations, like Ghana or Dominica, to reduce emissions there — and give Switzerland credit for it.
Here is an example of how it would work: Switzerland is paying to install efficient lighting and cleaner stoves in up to five million households in Ghana; these installations would help households move away from burning wood for cooking and rein in greenhouse gas emissions.
Then Switzerland, not Ghana, will get to count those emissions reductions as progress toward its climate goals.
Here is more from Hiroko Tabuchi at the NYT. Note that most of Swiss energy already is renewable, through hydroelectric and nuclear, yet of course there are many people complaining about this scheme.
4. James Steuart- overrated or underrated
5. Do border regions have above average cuisines (Sechuan, Bourgogne)
6. Zurich- overrated or underrated
Are there some great ethnic restaurants I haven’t heard of?
7.Would you have been a pagan or a Christian In the third century AD roman empire?
12. How would you compare the Swiss to the Irish enlightenment?
13. What is your Swiss take in general?
With numbers:
4. James Steuart was a Scottish economist and a precursor of Adam Smith, though more of a mercantilist. He had a good understanding of market structure, competition as a process of dynamic rivalry, and outlined how increasing competition would cause monopolistic prices to fall. He put forward a rudimentary understanding of supply and demand in his 1767 treatise. His macroeconomics is sometimes considered a precursor of Keynes, as it was demand side-oriented. He also put forward the idea of a purely abstract unit of account. Yet he isn’t talked about much, so definitely underrated.
5. Do border regions have better food? What exactly counts as a border region? The parts of the United States near Canada? The best food in Italy is not obviously at the (rather skimpy) borders. China and India might be the best food countries in the world, but because they are so large most of their cuisine is not “border cuisine.” So I say no.
6. Zurich is underrated, by everyone and everything, except the market prices for the real estate. The underrated side of the city includes the art scene (most of all the Kunsthaus), good ethnic food in the nooks and crannies, proximity to many good Swiss and German opera houses, proximity to the southern Schwarzwald, proximity to Basel, and proximity to Wallensee and many other wonderful short drives in Switzerland.
7. I still have the choice to be either a pagan or a Christian, and I am neither. Perhaps the same would be true for 3rd century AD Roman Empire Tyler.
12. I see the Swiss Enlightenment as centered in Albrecht von Haller, a mid-18th century poet (Die Alpen), scientist, polymath, and naturalist. I see the Swiss Enlightenment as focusing on two themes: a) coming to terms with a naturalistic rather than theological understanding of the beauties and world around them, and b) constructing an idealized narrative about Switzerland itself and its history and lifestyles. See also Salomon Gessner and Johann Jakob Bodmer. Can you count the Rousseau of this period as Swiss? Geneva had not yet joined the Swiss Confederation, but that is possibly another angle. How about the influence of Switzerland on Edward Gibbon?
13. My Swiss take in general is that the country and its success is radically understudied by outsiders.
