Results for “prediction market” 326 found
Tuesday assorted links
Holden Karnofsky emails me on transformative AI
Here is Holden, our discussion started with this post of mine, for his words I will use quotation marks rather than dealing with double indentation:
“…debates about specifics between climate scientists get incredibly intricate (and are often very sensitive to parameters we just can’t reasonably estimate), and if you tried to get oriented to climate science by reading one it would be a nightmare, but this doesn’t mean the big-picture ways in which climatologists diverge from conventional wisdom should be discounted.
I think the broad-brush picture here is a better starting point than an exchange between Eliezer, Ajeya, me and Scott.
Even shorter version:
- You can run the bio anchors analysis in a lot of different ways, but they all point to transformative AI this century;
- As do the expert surveys, as does Metaculus;
- Eliezer’s argument is that he thinks it will be sooner;
- The most naive extrapolations of economic growth trends imply singularity (or at least “new growth mode”) this century;
- Other angles of analysis (including the very-outside-view semi-informative priors) are basically about rebutting the idea that there’s a giant burden of proof here.
- Specific arguments for “later than 2100,” including outside-view arguments, seem reasonably close to nonexistent; Robin Hanson has a (unconvincing IMO) case for synthetic AI taking longer, but Robin is also forecasting transformative AI of a sort (ems, which he says will lead to an explosion in economic growth and a relatively quick transition to something even stranger) this century.
So I ultimately don’t see how you get under P=1/3 or so for this century, and if you are way under P=1/3, I’d be interested if there were any more you could say about why (though recognize forecasts can’t always totally be explained).
P=1/3 would put “transformative AI this century” within 2x of “nuclear war this century,” and I think the average “nuclear war” is way less likely (like at least 10x) to have super-long-run impacts than the average “transformative AI is developed.”
That’s my basic thinking! It’s based on numerous angles and is not very sensitive to specific takes on the rate at which FLOPs get cheaper, although at some point I hope we can nail that parameter down better via prediction markets or something of the sort. Prediction markets on transformative AI itself are going to be harder, but I’m hopeful about that too. I think a very fast transition is plausible, so it could be very bad news if folks like you continue thinking it’s a remote possibility until it’s obviously upon us. (In my analogy, today might be like early January was for COVID. We don’t know enough to be sure, but we know enough to be highly alert, and we won’t necessarily be sure very long before it’s too late.)”
End of Holden, now back to TC. And here is Holden’s “most important century” page. That is our century, people! This is all a bit of a follow-up on an in-person dialogue we had, but I will give him the last word (for now).
Saturday assorted links
1. New essays on the rise and fall of Swedish education, open access.
2. Summer courses at University of Austin. Instructors include Niall Ferguson, Ayaan Hirsi Ali, Rob Henderson, Kathleen Stock, Dorian Abbot, Deirdre McCloskey, David Mamet, Bari Weiss, Peter Boghossian, Joe Lonsdale, Arthur Brooks, Nadine Strossen, and Carlos Carvalho.
3. Lessons from shipbuilding productivity.
4. New academic prediction market, also covering matters Ukraine.
6. U.S. cutting back on purchases of Pfizer’s anti-Covid pill — crazy!
7. Robert Service has a good understanding of what is going on, and of Putin (WSJ).
8. China now getting antsy too, on Taiwan.
9. Good thread on Russia-China sanctions: “While China is Russia’s largest trading partner, but Russia is not even in China’s top ten.”
How to figure out where crypto is headed
That is the topic of my latest Bloomberg column, the piece has a number of ideas. You can start with this:
…the concept of relevance is focality, by which I mean the part of the system at which consumers direct their attention. Focality could determine whether crypto ushers in an era of dystopian inequality, or whether most of its benefits accrue to broader society.
That all sounds quite abstract, so consider a simple example from the world of music. Famous artists such as the Beatles or Taylor Swift attract attention with their very names — in other words, they have become focal. Then there are performance spaces or bars that are known for putting on good music, such as the Blue Note or, in an earlier era, the Fillmore. In this case, the venue is focal.
So the question is this: When people patronize crypto institutions, will they attach significance to the “innovator” or to the “intermediary”? Or, to continue the analogy with the music industry, the artist or the venue.
One scenario is that ordinary Americans will simply find crypto too confusing to deal with directly. Rather than choosing their favorite crypto assets, DeFi investments and NFT providers, they will outsource their decisions to well-known intermediaries. Imagine entering into a crypto contract with a company you have an established relationship with, such as a social media company, your bank or perhaps your labor union. The intermediary would deliver a “crypto package,” tailored to the needs of a broad swath of customers.
Significant parts of the crypto world would be relatively centralized….
I think you can imagine which problems would arise in that scenario, including the reemergence of de facto censorship. Alternately:
Another very different scenario: Users focus their attention on the crypto assets themselves, such as Bitcoin, Ether or Dogecoin. That kind of user focus would mean many of the gains of crypto accrue to the early crypto asset holders. Intermediaries (e.g., Coinbase) can earn a return, but the real brand name value would be held by the crypto asset itself.
Much of today’s crypto world looks like this, though it may not last as crypto broadens in applications and use. If you are long current crypto assets, you may be hoping for this kind of scenario to extend itself, because those assets will accumulate much of the value from higher crypto demand.
Yet another scenario: What if the attention of consumers were focused on the crypto innovators, who in this case would be analogous to better-known musical artists? One person may think “I like the DeFi options at Uniswap,” while another may say, “I am going to use the prediction markets over at Hedgehog.” In this scenario there is relatively little intermediation and heavy competition for consumer attention. Thus most of the gains from competition accrue to the users.
Customers would use or own or invest in crypto in a variety of ways, just as they listen to music on LPs, CDs, MP3s and streaming services. And in the same way that people share their playlists, crypto users could issue their own tokens (currencies) if they wanted, or serve as their own banks in the sense of making their own lending decisions and executing them autonomously.
I don’t know if people are up to all this work (or is it fun?). But in my view this is the best-case scenario — and the most technologically ambitious. Interestingly, crypto’s radical ability to disintermediate, if extended to its logical conclusion, could bring about a radical equalization of power that would lower the prices and values of the currently well-established crypto assets, companies and platforms.
So you can be bullish on crypto’s future without being bullish on current crypto prices. For a simple analogy, Spotify and YouTube have greatly expanded music’s reach, but overall the price of recorded music has fallen, and many performers earn much less than did their peers in the LP era. Or consider the agriculture sector, defined broadly: It has done very well over the last few centuries, but food prices have fallen rather than risen, due to higher output and greater competition.
Recommended.
Saturday assorted links
1. New Kalshi prediction markets in beta form.
2. The Austin Vernon blog. Or better link here.
3. The game theory of the infrastructure bill.
4. Should doctors be on first name terms? And U.S. rules that Bezos and Branson are not “astronauts.”
5. Using lottery winner results to predict that effects of UBI.
6. Interview in Spanish about Covid issues. And this far into the pandemic and the CDC still is systematically failing us.
Thursday assorted links
1. Skeptical results on prediction markets.
2. Jasper Johns owns some pretty amazing small Cezanne watercolors (scroll through to see them all).
3. Did urbanization come first?
4. Distributed teams in academia.
Where is (non-state capacity) libertarianism evolving?
That is the topic of my latest Bloomberg column, here is one bit:
I would say that the purer forms of libertarianism are evolving: from a set of policy stances on political questions to a series of projects for building entire new political worlds…
Much of the intellectual effort in libertarian circles is concentrated in two ideas in particular: charter cities and cryptocurrency.
Very recently a “charter city” was inaugurated in Honduras, with its own set of laws and constitutions, designed to set off an economic boom. Entrepreneurs are seeking to create such cities around the globe, typically as enclaves within established political units. The expectation is not that these cities would reflect libertarian doctrine in every way, but rather that they would be an improvement over prevailing governance, just as Hong Kong had much better outcomes than did Mao’s China.
A milder version of the charter cities concept is the YIMBY (“Yes In My Backyard”) movement, which is not founding new cities but seeking to transform existing ones by deregulating zoning and construction and thus building them out to a much greater extent.
Another area attracting energetic young talent is cryptocurrency. Bitcoin gets a lot of the attention, but it is a static system. The Ethereum project, led by Vitalik Buterin, is more ambitious. It is trying to create a new currency, legal system, and set of protocols for new economies on blockchains.
Unlike Bitcoin, Ethereum can be managed to better suit market demands. Imagine a future in which prediction markets are everywhere, micropayments are easy, self-executing smart contracts are a normal part of business, consumers own their own data and trade it on blockchains, and social media are decentralized and you can’t be canceled. The very foundations of banking and finance might move into this new realm.
Consistent with these developments, the most influential current figures in libertarianism have a strong background as doers: Elon Musk, Peter Thiel, Buterin and Balaji Srinivasan, to name a few, though probably none would qualify as a formal libertarian. All of them have strong roots outside the U.S., which perhaps liberated them from the policy debates that preoccupied American libertarians for so long.
The piece is 1200 words or so, 50% beyond the usual, plenty more at the link.
Thursday assorted links
1. “Work on things that aren’t prestigious.”
2. Patricia Lockwood on Ferrante.
3. How to get cancelled in Iceland, alternatively what not to look for in your supermarket’s PR director.
4. What do jobless men do all day? And the undermotivated apostate.
5. Vitalik on prediction markets, smart contract risk, epistemic humility, and more. Like most Vitalik, it is too good to excerpt.
6. Ezra Klein on work and child allowances (NYT). I agree with some but not all of this, in any case it is already obvious how much Ezra is in the very top tier of NYT columnists after only a few pieces. Every part of it is an actual argument, supported by evidence of some kind or another.
Wednesday assorted links
1. Arnold Kling on feminized culture.
2. SSC on vitamin D. He is fairly skeptical, I am more skeptical yet. The macro correlations that are there could be the result of many different forces, there is not much reason in theory to attribute such power to vitamin D.
3. How Paul Graham chose what to work on, painting, RISD, why he quit YC, and other stuff too.
4. Human challenge trials coming to the UK. And a brief but important comment: “This is most important experiment on COVID not yet done anywhere in the world. Can give us badly-needed data points on viral load, transmission and infection progression. Can later be used for vastly accelerated trials for vaccines, therapeutics and preventative approaches.”
5. Ben Southwood on how to build strong suburbs.
6. Cowen’s Second Law: “Accuracy of Urologic Conditions Portrayed on Grey’s Anatomy.”
7. Kalshi: real money prediction markets coming later this spring (WSJ).
Paul Milgrom, Nobel Laureate
Most of all this is a game theory prize and an economics of information prize, including game theory and asymmetric information. Much of the work has had applications to auctions and finance. Basically Milgrom was the most important theorist of the 1980s, during the high point of economic theory and its influence.
Here is Milgrom’s (very useful and detailed) Wikipedia page. Most of his career he has been associated with Stanford University, with one stint at Yale for a few years. Here is Milgrom on scholar.google.com. A very good choice and widely anticipated, in the best sense of that term. Here is his YouTube presence. Here is his home page.
Milgrom, working with Nancy Stokey, developed what is called the “no trade” theorem, namely the conditions under which market participants will not wish to trade with each other. Obviously if someone wants to trade with you, you have to wonder — what does he/she know that I do not? Under most reasonable assumptions, it is hard to generate a high level of trading volume, and that has remained a puzzle in theories of finance and asset pricing. People are still working on this problem, and of course it relates to work by Nobel Laureate Robert Aumann on when people should rationally disagree with each other.
Building on this no-trade result, Milgrom wrote a seminal piece with Lawrence Glosten on bid-ask spread. What determines bid-ask spread in securities markets? It is the risk that the person you are trading with might know more than you do. You will trade with them only when the price is somewhat more advantageous to you, so markets with higher degrees of asymmetric information will have higher bid-ask spreads. This is Milgrom’s most widely cited paper and it is personally my favorite piece of his, it had a real impact on me when I read it. You can see that the themes of common knowledge and asymmetric information, so important for the auctions work, already are rampant.
Alex will tell you more about auctions, but Milgrom working with Wilson has designed some auctions in a significant way, see Wikipedia:
Milgrom and his thesis advisor Robert B. Wilson designed the auction protocol the FCC uses to determine which phone company gets what cellular frequencies. Milgrom also led the team that designed the 2016-17 incentive auction, which was a two-sided auction to reallocate radio frequencies from TV broadcast to wireless broadband uses.
Here is Milgrom’s 277-page book on putting auction theory to practical use. Here is his highly readable JEP survey article on auctions and bidding, for an introduction to Milgrom’s prize maybe start there?
Here is Milgrom’s main theoretical piece on auctions, dating from Econometrica 1982 and co-authored with Robert J. Weber. it compared the revenue properties of different auctions and showed that under risk-neutrality a second-price auction would yield the highest price. Also returning to the theme of imperfect information and bid-ask spread, it showed that an expert appraisal would make bidders more eager to bid and thus raise the expected price. I think of Milgrom’s work as having very consistent strands.
With Bengt Holmstrom, also a Nobel winner, Milgrom wrote on principal-agent theory with multiple tasks, basically trying to explain why explicit workplace incentives and bonuses are not used more widely. Simple linear incentives can be optimal because they do not distort the allocation of effort across tasks so much, and it turned out that the multi-task principal agent problem was quite different from the single-task problem.
People used to think that John Roberts would be a co-winner, based on the famous Milgrom and Roberts paper on entry deterrence. Basically incumbent monopolists can signal their cost advantage by making costly choices and thereby scare away potential entrants. And the incumbent wishes to be tough with early entrants to signal to later entrants that they better had stay away. In essence, this paper was viewed as a major rebuttal to the Chicago School economists, who had argued that predatory behavior from incumbents typically was costly, irratoinal, and would not persist.
The absence of Roberts’s name on this award indicates a nudge in the direction of auction design and away from game theory a bit — the Nobel Committee just loves mechanism design!
That said, it is worth noting that the work of Milgrom and co-authors intellectually dominated the 1980s and can be identified with the peak of influence of game theory at that period of time. (Since then empirical economics has become more prominent in relative terms.)
Milgrom and Roberts also published a once-famous paper on supermodular games in 1990. I’ve never read it, but I think it has something to do with the possible bounding of strategies in complex settings, but based on general principles. This was in turn an attempt to make game theory more general. I am not sure it succeeded.
Milgrom and Roberts also produced a well-known paper finding the possible equilibria in a signaling model of advertising.
Milgrom and Roberts also wrote a series of papers on rent-seeking and “influence activities” within firms. It always seemed to me this was his underrated work and it deserved more attention. Among other things, this work shows how hard it is to limit internal rent-seeking by financial incentives (which in fact can make the problem worse), and you will see this relates to Milgrom’s broader work on multi-task principal-agent problems.
Milgrom also has a famous paper with Kreps, Wilson, and Roberts, so maybe Kreps isn’t going to win either. They show how a multi-period prisoner’s dilemma might sustain cooperating rather than “Finking” if there is asymmetric information about types and behavior. This paper increased estimates of the stability of tit-for-tat strategies, if only because with uncertainty you might end up in a highly rewarding loop of ongoing cooperation. This combination of authors is referred to as the “Gang of Four,” given their common interests at the time and some common ties to Stanford. You will note it is really Milgrom (and co-authors) who put Stanford economics on the map, following on the Kenneth Arrow era (when Stanford was not quite yet a truly top department).
Not what he is famous for, but here is Milgrom’s paper with Roberts trying to rationalize some of the key features of modern manufacturing. If nothing else, this shows the breadth of his interests and how he tries to apply game theory generally. One question they consider is why modern manufacturing has moved so strongly in the direction of greater flexibility.
Milgrom also has a 1990 piece with North and Weingast on the medieval merchant guilds and the economics of reputation, showing his more applied side. In essence the Law Merchant served as a multilateral reputation mechanism and enforced cooperation. Here is a 1994 follow-up. This work paved the way for later work by Avner Greif on related themes.
Another undervalued Milgrom piece is with Sharon Oster (mother of Emily Oster), or try this link for it. Here is the abstract:
The Invisibility Hypothesis holds that the job skills of disadvantaged workers are not easily discovered by potential new employers, but that promotion enhances visibility and alleviates this problem. Then, at a competitive labor market equilibrium, firms profit by hiding talented disadvantaged workers in low-level jobs.Consequently, those workers are paid less on average and promoted less often than others with the same education and ability. As a result of the inefficient and discriminatory wage and promotion policies, disadvantaged workers experience lower returns to investments in human capital than other workers.
With multiple, prestigious co-authors he has written in favor of prediction markets.
He was the doctoral advisor of Susan Athey, and in Alex’s post you can read about his auction advising and the companies he has started.
His wife, Eva Meyersson Milgrom, is herself a renowned social scientist and sociologist, and he met her in 1996 while seated next to her at a Nobel Prize dinner in Stockholm. Here is one of his papers with her (and Ravi Singh), on whether firms should share control with outsiders. Here is the story of their courtship.
Sunday assorted links
1. Emily Owens survey on the economics of policing. Or try this link.
2. Europe’s reopenings have mostly gone OK.
3. Give masks to people held in custody.
4. New policing bill introduced (NYT, seems mostly good).
5. Is cell-mediated immunity relevant? Important if true, speculative, but to my amateur mind increasingly not crazy. It is lacking in sufficient direct support to say “I believe it,” and needs to be tested against concrete alternatives at more micro levels, and it has some uncomfortable “residual” properties (hi Bob!), but so far it is the only theory that fits the data and surely that is worth something.
My avian flu blog days
Circa 2004 or so, it seemed to me that America was grossly underprepared for a possible pandemic. I started reading up on the topic, and I produced a very basic, simple Mercatus policy paper on avian flu. For obvious reasons, much of it is out of date and some of the recommendations have been adopted, but here is the first part of the Executive Summary:
1. The single most important thing we can do for a pandemic—whether avian flu or not—is to have well-prepared local health care systems. We should prepare for pandemics in ways that are politically sustainable and remain useful even if an avian flu pandemic does not occur.
2. Prepare social norms and emergency procedures which would limit or delay the spread of a pandemic. Regular hand washing, and other beneficial public customs, may save more lives than a Tamiflu stockpile.
3. Decentralize our supplies of anti-virals and treat timely distribution as more important than simply creating a stockpile.
4. Institute prizes for effective vaccines and relax liability laws for vaccine makers. Our government has been discouraging what it should be encouraging.
5. Respect intellectual property by buying the relevant drugs and vaccines at fair prices. Confiscating property rights would reduce the incentive for innovation the next time around.
6. Make economic preparations to ensure the continuity of food and power supplies. The relevant “choke points” may include the check clearing system and the use of mass transit to deliver food supply workers to their jobs.
7. Realize that the federal government will be largely powerless in the worst stages of a pandemic and make appropriate local plans.
8. Encourage the formation of prediction markets in an avian flu pandemic. This will give us a better idea of the probability of widespread human-to-human transmission.
9. Provide incentives for Asian countries to improve their surveillance. Tie foreign aid to the receipt of useful information about the progress of avian flu.
10. Reform the World Health Organization and give it greater autonomy from its government funders.
And also from later on:
4. We should not expect to choke off a pandemic in its country of origin. Once a pandemic has started abroad, we should shut schools and many public places immediately.
5. We should not obsess over avian flu at the expense of other medical issues. The next pandemic or public health crisis could come from any number of sources. By focusing on local preparedness and decentralized responses, this plan is robust to surprise and will also prove useful for responding to terrorism or natural catastrophes.
Still relevant today. For a while I also wrote an avian flu blog with Silviu Dochia, archived here.
Recession risk is rising
2020 @PredictIt recession prediction market probabilities are now above 40% amid #Coronavirus concerns.
That is from Jon Hartley, and here is his closely related new paper “Recession Prediction Markets and Macroeconomic Risk in Asset Prices.” Here is the prediction market page, at 47 as I am writing.
Tuesday assorted links
1. Felt anger in the previous day. A national ranking.
2. Which Medicare innovations really worked? (NYT)
3. Prediction markets in which research will replicate.
5. Black women in economics (NYT).
My Conversation with Hal Varian
Hal of course was in top form, here is the audio and transcript. Excerpt:
COWEN: Why doesn’t business use more prediction markets? They would seem to make sense, right? Bet on ideas. Aggregate information. We’ve all read Hayek.
VARIAN: Right. And we had a prediction market. I’ll tell you the problem with it. The problem is, the things that we really wanted to get a probability assessment on were things that were so sensitive that we thought we would violate the SEC rules on insider knowledge because, if a small group of people knows about some acquisition or something like that, there is a secret among this small group.
You might like to have a probability assessment of whether that would go through. But then, anybody who looks at the auction is now an insider. So there’s a problem in you have to find things that (a) are of interest to the company but (b) do not reveal financially critical information. That’s not so easy to do.
COWEN: But there are plenty of times when insider trading is either illegal or not enforced. Plenty of countries where it’s been legal, and there we don’t see many prediction markets in companies, if any. So it seems like it ought to have to be some more general explanation, or no?
VARIAN: Well, I’m just referring to our particular case. There was another example at the same time: Ford was running a market, and Ford would have futures markets on the price of gasoline, which was very relevant to them. It was an external price and so on. And it extended beyond the usual futures market.
That’s the other thing. You’re not going to get anywhere if you’re just duplicating a market that already exists. You have to add something to it to make it attractive to insiders.
So we ran a number of cases internally. We found some interesting behavior. There’s an article by Bo Cowgill on our experience with this auction. But ultimately, we ran into this problem that I described. The most valuable predictions would be the most sensitive predictions, and you didn’t want to do that in public.
And:
COWEN: But then you must think we’re not doing enough theory today. Or do you think it’s simply exhausted for a while?
VARIAN: Well, one area of theory that I’ve found very exciting is algorithmic mechanism design. With algorithmic mechanism design, it’s a combination of computer science and economics.
The idea is, you take the economic model, and you bring in computational costs, or show me an algorithm that actually solves that maximization problem. Then on the other side, the computer side, you build incentives into the algorithms. So if multiple people are using, let’s say, some communications protocol, you want them all to have the right incentives to have the efficient use of that protocol.
So that’s a case where it really has very strong real-world applications to doing this — everything from telecommunications to AdWords auctions.
And:
VARIAN: Yeah. I would like to separate the blockchain from just cryptographic protocols in general. There’s a huge demand for various kinds of cryptography.
Blockchain seems to be, by its nature, relatively inefficient. As an economist, I don’t like this proof of work that this is. I don’t like the fact that there’s one version of the blockchain that has to keep being updated. I don’t like the fact that it’s so slow. There are lots of things that you could fix, and I expect to see them fixed in the future, but I would say, crypto in general — big deal. Blockchain — not so much.
And finally:
COWEN: Now, users seem to like them both, but if I just look at the critics, why does it seem to me that Facebook is more hated than Google?
VARIAN: Well, you know, I actually don’t use Facebook. I don’t have any moral objection to it. I just don’t have the time to do it. [laughs] There are other things of this sort that can end up soaking up a substantial amount of time.
I think that one of the reasons — and this is, of course, quite speculative — I think that one of the reasons people are most worried about Facebook is they don’t really understand the limits of what can be done at Facebook. Whereas at Google, I think we’re pretty clear that we’re showing you ads. We’re showing you ads that are targeted to one thing or another, but that’s how the information’s used.
So, you’ve got this specific application in our case. In Facebook’s case, it’s more amorphous, I think.
There is much, much more at the link.