Marginal Revolution https://marginalrevolution.com/ Small Steps Toward A Much Better World Thu, 24 Sep 2026 13:53:45 +0000 en-US hourly 1 https://marginalrevolution.com/wp-content/uploads/2015/10/cropped-MR-logo-thumbnail-32x32.png Marginal REVOLUTION https://marginalrevolution.com/ 32 32 42693868 https://marginalrevolution.com/marginalrevolution/2026/09/thursday-assorted-links-571.html?utm_source=rss&utm_medium=rss&utm_campaign=thursday-assorted-links-571 Thursday assorted links https://feeds.feedblitz.com/~/969897881/0/marginalrevolution~Thursday-assorted-links.html https://marginalrevolution.com/marginalrevolution/2026/09/thursday-assorted-links-571.html#comments Thu, 24 Sep 2026 16:07:24 +0000 https://marginalrevolution.com/?p=93871 1. Shruti on Coltrane. 2. On the drop in Pisa scores.  Quite exaggerated, it turns out. 3. Hope you all reoptimize your views accordingly, on AI agreements with China. 4. Dutch economy is picking up. 5. One professor’s AI-assisted output.  Hey people, solve for the equilibrium… 6. Brisket is now too expensive for this Texas […]

The post Thursday assorted links appeared first on Marginal REVOLUTION.

]]> 1. Shruti on Coltrane.

2. On the drop in Pisa scores.  Quite exaggerated, it turns out.

3. Hope you all reoptimize your views accordingly, on AI agreements with China.

4. Dutch economy is picking up.

5. One professor’s AI-assisted output.  Hey people, solve for the equilibrium…

6. Brisket is now too expensive for this Texas smokehouse.

7. I asked the new Claude to write a 900-word column on refineries, supply, and how to improve the problem.

The post Thursday assorted links appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/thursday-assorted-links-571.html/feed 35 93871 https://marginalrevolution.com/marginalrevolution/2026/09/ca-91326-a-public-choice-mystery-investigation.html?utm_source=rss&utm_medium=rss&utm_campaign=ca-91326-a-public-choice-mystery-investigation CA 91326: A Public Choice Mystery Investigation https://feeds.feedblitz.com/~/969821807/0/marginalrevolution~CA-A-Public-Choice-Mystery-Investigation.html https://marginalrevolution.com/marginalrevolution/2026/09/ca-91326-a-public-choice-mystery-investigation.html#comments Thu, 24 Sep 2026 11:15:32 +0000 https://marginalrevolution.com/?p=93848 Carl Danner writes to me with a political economy mystery. On Sept 9, 2026, The California Public Utilities Commission announced a set of clean energy grants but only “for eligible organizations headquartered in ZIP Code 91326.” I decided to investigate with ChatGPT. The first thing we discovered is that: The ZIP-code restriction is about the […]

The post CA 91326: A Public Choice Mystery Investigation appeared first on Marginal REVOLUTION.

]]> Carl Danner writes to me with a political economy mystery. On Sept 9, 2026, The California Public Utilities Commission announced a set of clean energy grants but only “for eligible organizations headquartered in ZIP Code 91326.”

I decided to investigate with ChatGPT. The first thing we discovered is that:

The ZIP-code restriction is about the Aliso Canyon gas leak. More specifically, this is money ultimately intended to benefit residents of Porter Ranch and nearby communities affected by the 2015–16 Aliso Canyon disaster, not a generic clean-energy grant that happens to be located there.

The Aliso Canyon gas leak was a big deal with thousands of families evacuated for months so it’s not surprising that money is allocated to that problem but note that the grants are not for organization serving the people or region affected by the gas leak (which is much larger than zip 91326) but for an organization headquartered in ZIP Code 91326. I smelled a payoff. So I pushed ChatGPT to investigate:

It may be [a payoff], but the evidence supports a narrower claim: the Round 3 rules look unusually tailored to make Save Porter Ranch a likely recipient.

The strongest facts are:

  • The legislature appropriated $2 million from the Aliso Canyon Recovery Account for CBO outreach in Los Angeles County, prioritizing a broad “Aliso Canyon Disaster Area” containing 11 communities—not Porter Ranch alone.
  • CPUC’s first round distributed roughly $1.3 million through a normal competitive process across multiple organizations.
  • For Round 3, CPUC suddenly restricts eligibility to organizations headquartered in the single ZIP code 91326, with a maximum award of $75,000 and exclusion of previous recipients.
  • Save Porter Ranch is headquartered at 19360 Rinaldi Street, Porter Ranch, CA 91326 and has been an active participant in CPUC proceedings concerning Aliso Canyon for years.

That is suggestive because the legislative authorization itself does not appear to require this hyper-local 91326 restriction. It defines the relevant disaster area much more broadly. CPUC therefore seems to have exercised administrative discretion to create a tiny eligibility pool.

I would distinguish three possibilities:

  1. Benign geographic set-aside: CPUC noticed that none of the first-round money had gone to an organization actually based in Porter Ranch and reserved the final $75,000 for a local group.
  2. De facto earmark: CPUC knew that Save Porter Ranch was one of very few qualifying CBOs in 91326 and wrote the eligibility rule sufficiently narrowly that it was highly likely to win.
  3. Quid pro quo/payoff: the money was deliberately directed to Save Porter Ranch in exchange for support, silence, settlement behavior, etc.

We have substantial circumstantial evidence for #2, but I have seen no evidence yet for #3. Calling it a “payoff” requires some evidence about why CPUC created the ZIP restriction or about communications with Save Porter Ranch.

I wasn’t satisfied so I pushed the investigation deeper and ChatGPT investigated the legislative history. It found that

The 91326 restriction did not originate with CPUC staff. The Legislature itself inserted it into the 2026 budget cleanup bill, and it is unusually specific….The evidence [for a payoff] is considerably stronger than I thought.

…If I were writing this up now, the sharp factual formulation would be: “In an August 2026 budget bill, the California Legislature quietly directed the CPUC to hold an otherwise-unplanned third grant round restricted to nonprofits headquartered in a single Porter Ranch ZIP code. One conspicuous beneficiary of that restriction is Save Porter Ranch, the local Aliso Canyon advocacy organization, which had previously helped lobby state officials over how Aliso Canyon penalty funds should be spent.”

I then asked ChatGPT to “look for overlap in Save Porter Ranch and CA legislatures, e.g. spouses, children, business partners.” It replied:

The strongest link is Richard Mathews. Save Porter Ranch’s IRS filings list Mathews as a director alongside Matt Pakucko and Kyoko Hibino. Mathews is also the longtime president of the North Valley Democratic Club. He previously served as a regional vice chair of the Los Angeles County Democratic Party and on the California Democratic Party executive board and its Legislation Committee.

That puts a Save Porter Ranch director directly inside the local/state Democratic Party apparatus, not merely in occasional contact with elected officials.

…the Mathews connection is stronger than I expected. This isn’t just an advocacy organization obtaining support from politicians. One of Save Porter Ranch’s three disclosed directors was simultaneously a fairly senior Democratic Party insider whose political club includes and supports legislators involved in the relevant legislative machinery.

…The missing piece remains the most important one: who requested the sentence “headquartered within ZIP Code 91326”?

I enjoyed the fact that ChatGPT seems excited by the new discoveries! Digging deeper:

That gives us a much tighter network than we had initially:

Save Porter Ranch
→ director Richard Mathews
→ president, North Valley Democratic Club
→ Schiavo and Gabriel are identified as club members / repeatedly endorsed by it
→ Schiavo sits on Budget + Utilities & Energy and represents the affected area
→ Gabriel chairs Assembly Budget
→ August 28 Assembly budget amendments insert a bespoke ZIP-code restriction
→ restriction makes Save Porter Ranch an unusually obvious potential recipient.

There is another revealing feature. The amendment was handled as part of a budget-bill junior rather than through ordinary policy legislation. The public legislative history is extraordinarily thin: amendment August 28, withdrawn from committee August 30, passed August 31. Neither Gabriel’s Assembly presentation nor Laird’s Senate presentation told legislators that the “third round” was actually limited to one ZIP code. Gabriel merely said CPUC must offer “a third round of unspent funds”; Laird similarly described a third round “as specified.” The Assembly passed it 72–0.

That doesn’t demonstrate corruption, but it is exactly how a highly localized legislative earmark can pass with effectively zero scrutiny.

Alas, “the remaining decisive evidence probably isn’t indexed on the web.” Still, you get the idea.

Of course, this is small potatoes by the standards of California political/regulatory friends and families but it’s a nice case study in political economy. How do we want government to work? Is this ethical? What if the sums were much larger? What if the funds went to an equally well-connected group but for a less justifiable cause?   

It’s always surprised me that most public choice is not done in this micro-history way. My early paper on the Separation of Commercial and Investment Banking: The Morgans vs. The Rockefellers (which was inspired by Burch’s magnificent Elites in American History which hardly any public choice scholars know) is one of the few examples. I could have done many more papers along those lines but archival history didn’t appeal to me although I did learn a lot digging through Carter Glass’s papers at UVA. I suppose economists like models and consider this sort of thing journalism or history at best.

If you want more details on CA 91326 you can read my whole conversation with ChatGPT.

The post CA 91326: A Public Choice Mystery Investigation appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/ca-91326-a-public-choice-mystery-investigation.html/feed 20 93848 https://marginalrevolution.com/marginalrevolution/2026/09/fear-of-data.html?utm_source=rss&utm_medium=rss&utm_campaign=fear-of-data *Fear of Data* https://feeds.feedblitz.com/~/969780677/0/marginalrevolution~Fear-of-Data.html https://marginalrevolution.com/marginalrevolution/2026/09/fear-of-data.html#comments Thu, 24 Sep 2026 08:24:05 +0000 https://marginalrevolution.com/?p=93864 The author is Omri Ben-Shahar, and the subtitle is How Privacy Panic Led Tech Regulation Astray — and How to Fix It.  I would describe this book as bracing, and full of substantive engagement.  Basically the author wishes to give privacy considerations less weight in social decisions.  Excerpt: What is the concrete evidence for the […]

The post *Fear of Data* appeared first on Marginal REVOLUTION.

]]> The author is Omri Ben-Shahar, and the subtitle is How Privacy Panic Led Tech Regulation Astray — and How to Fix It.  I would describe this book as bracing, and full of substantive engagement.  Basically the author wishes to give privacy considerations less weight in social decisions.  Excerpt:

What is the concrete evidence for the benefits of facial recognition technology in investigation of  human-trafficking crimes?  I would love to have found global estimates of the magnitude — of the trafficking victims rescued through the most advanced facial recognition methods — but all I have is a collage of reports [reports are then described].

One chapter is entitled “The Futility of Personal Rights.”  Agree or not, this book is full of actual arguments, so I approve.

The post *Fear of Data* appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/fear-of-data.html/feed 33 93864 https://marginalrevolution.com/marginalrevolution/2026/09/my-economics-rich-conversation-with-gita-gopinath.html?utm_source=rss&utm_medium=rss&utm_campaign=my-economics-rich-conversation-with-gita-gopinath My economics-rich Conversation with Gita Gopinath https://feeds.feedblitz.com/~/969730616/0/marginalrevolution~My-economicsrich-Conversation-with-Gita-Gopinath.html https://marginalrevolution.com/marginalrevolution/2026/09/my-economics-rich-conversation-with-gita-gopinath.html#comments Thu, 24 Sep 2026 04:41:41 +0000 https://marginalrevolution.com/?p=93872 Here is the audio, video, and transcript.  Here is the episode summary: Gita Gopinath has spent the last several years inside the room where the world’s monetary plumbing gets fixed — or doesn’t. As first deputy managing director of the IMF and, before that, the Fund’s chief economist, she worked on sovereign debt restructuring, followed […]

The post My economics-rich Conversation with Gita Gopinath appeared first on Marginal REVOLUTION.

]]> Here is the audio, video, and transcript.  Here is the episode summary:

Gita Gopinath has spent the last several years inside the room where the world’s monetary plumbing gets fixed — or doesn’t. As first deputy managing director of the IMF and, before that, the Fund’s chief economist, she worked on sovereign debt restructuring, followed Argentina’s disinflation up close, and built her academic reputation on ideas like dollar dominance and the dominant currency paradigm. Now back at Harvard launching a new Global Economics Lab, her conversation with Tyler ranges from why exchange rates don’t adjust the way models predict, to what stablecoins really cost, to whether AI’s growth dividend will show up on schedule.

Tyler and Gita discuss why trade balances are only weakly related to real exchange rates, before turning to Argentina, Milei’s fiscal discipline, the peso, and dollarization. Then they discuss whether economists got the euro wrong; whether trade imbalances with China are really the problem; stablecoins, CBDCs, and the Treasury market; and whether AI will ease the debt burden, cause stagflation, or transform economics itself. Finally, they turn to her family’s Kerala roots: why the state keeps electing communists, the relationship between matriarchy and social indicators, single-sex education, and what she’s building next.

Excerpt:

COWEN: Now, Scott Sumner argues we shouldn’t be worried about trade imbalances at all. We don’t worry about them across American states. A lot of the EU has a pretty big current account surplus. Not many people are worried about that. Yet when it comes to China, the talk is all about trade imbalances. Scott argues that’s more of a political project than an actual economic argument. Do you agree with Scott or not?

GOPINATH: I would agree with Scott along the following lines, which is that the trade imbalance in and of itself is not something that we should be focused on. I think what we care about is welfare, and welfare involves jobs and consumption, what’s inflation, purchasing power, and so on. People do not wake up in the morning saying, “Okay, my current account deficit is too big, or my current account surplus is too big.” If all of your policies that were delivering good outcomes for your country were to bring along a deficit or a surplus, that’s perfectly fine. There are lots of good reasons to be running deficits and surpluses we know. There’s nothing that tells you that you shouldn’t.

The problem arises when you have policies that countries have in place that are inconsistent with any kind of a balanced growth model, and they manifest themselves in that trade deficits and surpluses. This is what I’m saying as an economist.

Now, of course, from a politician’s perspective, there are other arguments for why they would point to it. What we have seen historically over and over again is that whenever you had these increases in these deficits and surpluses, you’ve had trade wars or calls for protectionism. That’s what happened during Reagan’s time, 1980s. That’s what led to the Plaza Accord and then all the adjustments that followed after that.

Then you also do worry about the possibility of crises. The Great Financial Crisis was preceded by growing imbalances. There was a sense in which all this big savings glut. All this money flushing around. All these large surpluses and deficits were part of the problem. Now I would say where we are now in this third wave of concern about imbalances.

To be clear, it’s not the imbalance itself. As an economist, I would say that it’s not the imbalance in itself. You don’t wake up and say, “This is what I’m trying to prevent. I’m trying to prevent us from having a deficit. I want us to have balanced trade.” I think that’s bad economics.

COWEN: Say China is channeling what would have been wage income into investment, and that’s plausibly the case. Now, it may be politically unstable in the sense that we Americans object to it, but that’s not an argument per se, right? We have a choice as to whether or not we should object to it. It doesn’t seem that rigorous to say, well, this will cause another financial crisis like 2008. There’s just not real evidence for that. Scott Sumner would say, “Let’s just be happy we have cheaper goods. We send them paper. We get back stuff and go our merry way.” Why is that wrong?

GOPINATH: A few things. Firstly, I could go into these differences between trade deficits and surpluses versus what I would call sectoral imbalances, or the fact that China’s running a big manufacturing surplus, which is different from a trade surplus itself, or that it has big EV production. What we do recognize, again, is that we do a pretty bad job in moving or helping workers that have lost their jobs in certain sectors.

COWEN: We’re at full employment now or very close to it.

GOPINATH: Again, we’re at full employment right now, but we’ve had this period of time when we had communities that were deeply affected by—not just trade—automation was a big part of it too. We don’t do a great job in terms of getting people back into jobs, and that can affect, depending upon your ideal welfare function for the country, if you care a lot about those people, then obviously in that case, that’s something that should matter in your policy decisions.

In the case of China, I would say firstly that, in both the case of the US and China, we’re very far from the world where countries are doing good policies, and this is all about comparative advantage, and this is all about this would be the outcome we would have in a world where every country was doing the right policy. We’re very far from that.

China’s surpluses are a reflection of things going wrong in China. It’s not a reflection of strength in China. It’s a reflection of weak consumption. It’s a reflection of misallocated resources going into different sectors. They did that with their property markets. Now they have a huge property market problem that they haven’t been able to fix in five years. They have this now with other markets, including EVs and the other sectors. They have the problem with inflation being too low.

They’re trying to do so called anti-involution policies, bringing companies together and telling them, “You’ve got to keep prices higher than what you’re doing right now.” This is not the world where they’re playing good policies, and this is all comparative advantage, and this is the outcome that we see. I wouldn’t push the argument that this is, “Oh, we should be just happier with cheaper goods from China.”

Economics throughout…

The post My economics-rich Conversation with Gita Gopinath appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/my-economics-rich-conversation-with-gita-gopinath.html/feed 22 93872 https://marginalrevolution.com/marginalrevolution/2026/09/from-the-master-of-industrial-organization.html?utm_source=rss&utm_medium=rss&utm_campaign=from-the-master-of-industrial-organization From the Master of Industrial Organization https://feeds.feedblitz.com/~/969660365/0/marginalrevolution~From-the-Master-of-Industrial-Organization.html https://marginalrevolution.com/marginalrevolution/2026/09/from-the-master-of-industrial-organization.html#comments Wed, 23 Sep 2026 21:40:20 +0000 https://marginalrevolution.com/?p=93874 Believing that AI will be incorrigible leads us to proposing the wrong solutions, including ones which are likely to increase the danger than decrease it. I think that an AI pause is using the wrong method to fix an ill-posed problem. The point of a pause would be to research AI until we have a […]

The post From the Master of Industrial Organization appeared first on Marginal REVOLUTION.

]]>

Believing that AI will be incorrigible leads us to proposing the wrong solutions, including ones which are likely to increase the danger than decrease it. I think that an AI pause is using the wrong method to fix an ill-posed problem. The point of a pause would be to research AI until we have a deep enough theoretical understanding as to predict what an AI will do. I think this goal will always elude us. I cannot conceive of what this deep theoretical understanding would look like; I don’t think anyone actually does. It’s a call for a stop, not a pause.

The world would not stand still, though. The companies far behind the frontier now would doubtless learn more about what it is that the frontier AI companies are doing. If cooperation were to break down, we would have restarted the race with many more competitors. This is no improvement.

Nicholas Decker throughout.

The post From the Master of Industrial Organization appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/from-the-master-of-industrial-organization.html/feed 21 93874 https://marginalrevolution.com/marginalrevolution/2026/09/optimal-liability-for-offensive-and-defensive-ai.html?utm_source=rss&utm_medium=rss&utm_campaign=optimal-liability-for-offensive-and-defensive-ai Optimal liability for offensive and defensive AI https://feeds.feedblitz.com/~/969622667/0/marginalrevolution~Optimal-liability-for-offensive-and-defensive-AI.html https://marginalrevolution.com/marginalrevolution/2026/09/optimal-liability-for-offensive-and-defensive-ai.html#comments Wed, 23 Sep 2026 18:03:36 +0000 https://marginalrevolution.com/?p=93867 How much liability should AI providers bear when their services enable both attack and defence? Liability can improve welfare while increasing harm. Providers sell a common input to productive users, attackers and defenders. Within a defended contest, a higher common price reduces effort without changing attack success or attacker profits, saving resources and improving the […]

The post Optimal liability for offensive and defensive AI appeared first on Marginal REVOLUTION.

]]>

How much liability should AI providers bear when their services enable both attack and defence? Liability can improve welfare while increasing harm. Providers sell a common input to productive users, attackers and defenders. Within a defended contest, a higher common price reduces effort without changing attack success or attacker profits, saving resources and improving the target’s security payoff. Compensation weakens defence and raises attacker profits. Optimal liability balances these effects against productive exclusion. Greater competition can lower optimal liability; every such decline must end at an outcome retaining defence. With cybersecurity access fixed, monopoly can warrant partial liability but never full liability when provision is worthwhile. When guardrails preserving productive uses are available, strong competition favours universal guarding socially but encourages unilateral removal at insufficient liability. At a fixed provider count, sufficiently many productive users ensure a pure equilibrium with universal guarding under high liability. A universal-guarding requirement makes liability redundant. Under monopoly, adoption follows a unique liability threshold, while zero liability remains uniquely optimal for a range of parameters with sufficiently many productive users.

That is from a new paper by Joshua Gans.

The post Optimal liability for offensive and defensive AI appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/optimal-liability-for-offensive-and-defensive-ai.html/feed 24 93867 https://marginalrevolution.com/marginalrevolution/2026/09/wednesday-assorted-links-572.html?utm_source=rss&utm_medium=rss&utm_campaign=wednesday-assorted-links-572 Wednesday assorted links https://feeds.feedblitz.com/~/969601814/0/marginalrevolution~Wednesday-assorted-links.html https://marginalrevolution.com/marginalrevolution/2026/09/wednesday-assorted-links-572.html#comments Wed, 23 Sep 2026 15:48:52 +0000 https://marginalrevolution.com/?p=93866 1. Diving northern gannets. 2. Guatemala update. 3. Tech-utopian project coming in Uruguay? 4. “Toy sales among adult-only households have now outpaced households with children for toy sales, growing 16% through June, according to a market research group.” 5. Do stock buybacks bring harm? 6. How AI competition works in China. 7. Economist column on […]

The post Wednesday assorted links appeared first on Marginal REVOLUTION.

]]> 1. Diving northern gannets.

2. Guatemala update.

3. Tech-utopian project coming in Uruguay?

4. “Toy sales among adult-only households have now outpaced households with children for toy sales, growing 16% through June, according to a market research group.”

5. Do stock buybacks bring harm?

6. How AI competition works in China.

7. Economist column on AI risk and shorting the market (hint: ask the AGI what to do!  And of course asset prices fall along the way).

The post Wednesday assorted links appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/wednesday-assorted-links-572.html/feed 46 93866 https://marginalrevolution.com/marginalrevolution/2026/09/the-price-of-intelligence-is-falling.html?utm_source=rss&utm_medium=rss&utm_campaign=the-price-of-intelligence-is-falling The Price of Intelligence is Falling Rapidly https://feeds.feedblitz.com/~/969549917/0/marginalrevolution~The-Price-of-Intelligence-is-Falling-Rapidly.html https://marginalrevolution.com/marginalrevolution/2026/09/the-price-of-intelligence-is-falling.html#comments Wed, 23 Sep 2026 11:17:32 +0000 https://marginalrevolution.com/?p=93870 An amazing Epoch AI report from Emberson and Roodman: Over the past three years, the cost of a given level of AI performance has fallen an average of some 47% per quarter. That is a 13-fold drop every year – a faster rate than any other transformative technology in history. To give an example, OpenAI […]

The post The Price of Intelligence is Falling Rapidly appeared first on Marginal REVOLUTION.

]]> An amazing Epoch AI report from Emberson and Roodman:

  • Over the past three years, the cost of a given level of AI performance has fallen an average of some 47% per quarter. That is a 13-fold drop every year – a faster rate than any other transformative technology in history.

To give an example, OpenAI o3 cost about $0.30/question to attain 75% on GPQA Diamond in January 2025, while GPT-5.6 Luna attained roughly the same score for $0.0004 in mid-2026—a roughly 725-fold decline in under 18 months.

Thus, it’s not just that the models are getting smarter. A given level of intelligence is also requiring dramatically less inference expenditure. This is one reason the open-model threat is not as large as it appears: frontier models don’t merely outperform older models; they are rapidly becoming cheaper to run at any given level of performance. Smarter and cheaper.

The post The Price of Intelligence is Falling Rapidly appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/the-price-of-intelligence-is-falling.html/feed 65 93870 https://marginalrevolution.com/marginalrevolution/2026/09/what-ive-been-reading-295.html?utm_source=rss&utm_medium=rss&utm_campaign=what-ive-been-reading-295 What I’ve been reading https://feeds.feedblitz.com/~/969521099/0/marginalrevolution~What-Ive-been-reading.html https://marginalrevolution.com/marginalrevolution/2026/09/what-ive-been-reading-295.html#comments Wed, 23 Sep 2026 04:40:51 +0000 https://marginalrevolution.com/?p=93835 1. Constance Reid, Hilbert.  We should all be reading more mathematics these days, right?  Because the field is “on fire.”  Hilbert is not a bad place to start, and I found this book clear and biographical in a good way.  It is less on the math and more closely tied to broader German history.  Soon […]

The post What I’ve been reading appeared first on Marginal REVOLUTION.

]]> 1. Constance Reid, Hilbert.  We should all be reading more mathematics these days, right?  Because the field is “on fire.”  Hilbert is not a bad place to start, and I found this book clear and biographical in a good way.  It is less on the math and more closely tied to broader German history.  Soon to come is Jeremy J. Gray, The Hilbert Challenge.

2. Ben Buchanan and Tantum Collins, The Bitter Struggle: Superintelligence, Superpowers, and the Fate of the World.  This will become the standard source on the history of chip bans, the surrounding technological developments, and chip and AI policy under the Biden administration.  Written by the people who were there.

3. Jeanna Smialek, The Invisible Hand of Maria Edgeworth: How a Nineteenth-Century Novelist Taught the World Economics.  Maria Edgeworth was an Anglo-Irish writer and novelist, Castle Rackrent being my favorite book by her.  This study of her work is excellent and much-needed.

4. Anton Jäger Hyperpolitics: Extreme Politicization without Political Consequences.  I am sorry, but this book just isn’t very good.  The title and subtitle are spot on, but the author just does not have enough accurate factual knowledge about the world.

5. Robert Alter, The Language of Fiction: The Craft of Writing and the Pleasures of Reading.  A wonderful short take on why reading fiction is so worthwhile.  As one might expect, Alter has exquisite taste.  The last two pages of the work are marred by some not very good comments about LLMs, but overall this is excellent.

There is Ed Conway, Trade World: The Ties that Bound Our Past and May Unravel Our Future.

Owen Zidar and Eric Zwick, The Everywhere Millionaire: Who is Really Rich in America and How They Got There.  Not a b.s. whiner sort of book, based on real data and written by real economists.

Matthew Botvinick, AI & Political Freedom: The Risk to Democracy and How to Respond, contains plenty of analysis and social science.  The author is now with Anthropic.

And Branko Milanovic, The Great Global Transformation: The United States, China, and the Remaking of the World Economic Order.  I find his concept of National Market Liberalism useful.

The post What I’ve been reading appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/what-ive-been-reading-295.html/feed 25 93835 https://marginalrevolution.com/marginalrevolution/2026/09/the-economics-of-late-grandmotherhood.html?utm_source=rss&utm_medium=rss&utm_campaign=the-economics-of-late-grandmotherhood The economics of late grandmotherhood https://feeds.feedblitz.com/~/969496097/0/marginalrevolution~The-economics-of-late-grandmotherhood.html https://marginalrevolution.com/marginalrevolution/2026/09/the-economics-of-late-grandmotherhood.html#comments Tue, 22 Sep 2026 18:52:28 +0000 https://marginalrevolution.com/?p=93862 We show how fertility delays compound across generations to raise age at first grandmotherhood. Under the 1960 first-birth schedule median grandmotherhood came at 44. Under the 2024 schedule it is 57, cutting the healthy years with grandchildren from 28 to 15. Longer life does not make up the loss of years when the grandmother is […]

The post The economics of late grandmotherhood appeared first on Marginal REVOLUTION.

]]>

We show how fertility delays compound across generations to raise age at first grandmotherhood. Under the 1960 first-birth schedule median grandmotherhood came at 44. Under the 2024 schedule it is 57, cutting the healthy years with grandchildren from 28 to 15. Longer life does not make up the loss of years when the grandmother is healthy enough to supply money and care. Delay is most pronounced among the college-educated, with reproductive-control access accounting for over half the fall in grandmotherhood by age 55 among White college graduates. When a daughter times her first birth to maximize her own payoff, the loss her choice imposes on her mother is external to her calculation. We call this the Rotten Adult Kid Problem. Delay raises both the price of a grandmother’s care and wealth per grandchild.

That is from a new NBER working paper by Dora Costa and Matt Kahn.

The post The economics of late grandmotherhood appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/the-economics-of-late-grandmotherhood.html/feed 68 93862 https://marginalrevolution.com/marginalrevolution/2026/09/tuesday-assorted-links-587.html?utm_source=rss&utm_medium=rss&utm_campaign=tuesday-assorted-links-587 Tuesday assorted links https://feeds.feedblitz.com/~/969490691/0/marginalrevolution~Tuesday-assorted-links.html https://marginalrevolution.com/marginalrevolution/2026/09/tuesday-assorted-links-587.html#comments Tue, 22 Sep 2026 16:10:01 +0000 https://marginalrevolution.com/?p=93856 1. Designing a FINRA for frontier AI. 2. Andrew Ng. 3. Some semi-skeptical notes on stablecoins, based on economic history. 4. Global views on fairness and redistribution. 5. Essays on whether AI will make management more valuable. 6. The real authorities on the California billionaire wealth tax.  They are right in front of your eyes. […]

The post Tuesday assorted links appeared first on Marginal REVOLUTION.

]]> 1. Designing a FINRA for frontier AI.

2. Andrew Ng.

3. Some semi-skeptical notes on stablecoins, based on economic history.

4. Global views on fairness and redistribution.

5. Essays on whether AI will make management more valuable.

6. The real authorities on the California billionaire wealth tax.  They are right in front of your eyes.

7. “The storied venture-capital firm is investing $35 million in an unaccredited San Francisco-based two-year alternative to traditional college for students ages 16 to 22. Though named the Horowitz Andreessen Academy, the school will operate as an independent business led by Gagan Biyani, co-founder of the ed-tech platform Udemy.” (WSJ)

The post Tuesday assorted links appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/tuesday-assorted-links-587.html/feed 77 93856 https://marginalrevolution.com/marginalrevolution/2026/09/dartmouth-provost-okie-dokie.html?utm_source=rss&utm_medium=rss&utm_campaign=dartmouth-provost-okie-dokie Dartmouth Provost okie-dokie https://feeds.feedblitz.com/~/969472844/0/marginalrevolution~Dartmouth-Provost-okiedokie.html https://marginalrevolution.com/marginalrevolution/2026/09/dartmouth-provost-okie-dokie.html#comments Tue, 22 Sep 2026 12:20:22 +0000 https://marginalrevolution.com/?p=93863 On Aug. 26, a Semafor investigation using Pangram to determine whether recent guest columns in major newspapers were written using artificial intelligence identified a column by Schnell as “100% AI-written.” Schnell’s column in The Washington Post — titled “Universities are fighting AI cheating. But there’s a deeper problem.” — argued that “a degree should distinguish what students can do […]

The post Dartmouth Provost okie-dokie appeared first on Marginal REVOLUTION.

]]>

On Aug. 26, a Semafor investigation using Pangram to determine whether recent guest columns in major newspapers were written using artificial intelligence identified a column by Schnell as “100% AI-written.” Schnell’s column in The Washington Post — titled “Universities are fighting AI cheating. But there’s a deeper problem.” — argued that “a degree should distinguish what students can do independently from what they can accomplish with AI.

”The Dartmouth subsequently reviewed publications written by Schnell before and after the November 2022 public release of ChatGPT. Pangram 4.0, the latest publicly available version of the AI detector, labeled all of Schnell’s written works prior to 2022 as “100% human-written,” while a test of nine of his works published this year returned a median “AI-written” percentage of 96%. Between 2019 and 2026, Schnell did not publish any academic works or op-eds for which he was the sole author credited, according to his online curriculum vitae.Pangram 4.0 assesses whether a work is AI-written by separating the submitted text into smaller segments, which are then identified as “human,” “AI-assisted” or “AI-generated,” according to a Pangram technical report.

…Schnell’s Sept. 18 email statement to The Dartmouth was itself “100% AI-written,” according to a Pangram test.

Here is the full story, you people crack me up…

The post Dartmouth Provost okie-dokie appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/dartmouth-provost-okie-dokie.html/feed 75 93863 https://marginalrevolution.com/marginalrevolution/2026/09/banning-self-recursive-improvement-in-ai-models.html?utm_source=rss&utm_medium=rss&utm_campaign=banning-self-recursive-improvement-in-ai-models Banning self-recursive improvement in AI models? https://feeds.feedblitz.com/~/969457874/0/marginalrevolution~Banning-selfrecursive-improvement-in-AI-models.html https://marginalrevolution.com/marginalrevolution/2026/09/banning-self-recursive-improvement-in-ai-models.html#comments Tue, 22 Sep 2026 04:45:57 +0000 https://marginalrevolution.com/?p=93857 Some people are suggesting this, including Ezra Klein. I do not understand how it is supposed to work.  Put aside the issue of foregone innovations, let us say I seek to access an AI from overseas.  Am I allowed to visit their website?  So many sites have AI behind them, including Chinese open source AI.  […]

The post Banning self-recursive improvement in AI models? appeared first on Marginal REVOLUTION.

]]> Some people are suggesting this, including Ezra Klein.

I do not understand how it is supposed to work.  Put aside the issue of foregone innovations, let us say I seek to access an AI from overseas.  Am I allowed to visit their website?  So many sites have AI behind them, including Chinese open source AI.  Do all of those sites get banned?  And how?  Firewall imposed on Americans?  And we cannot put the good open source models on our hard drives?  How enforced?  Does the government have to read the creation logs of the model (how can they?), and then decide whether I can use that model or not?

What if Anthropic licenses its IP to an independent subsidiary in the Cayman Islands and they keep on using code to improve the AI?  Or maybe only foreign nations are allowed to have the best models and the really fast rates of improvement?  How are multinationals supposed to operate across borders?  (Presumably you cannot just use the RSI of your foreign affiliate, but then the whole MNC is crippled and eventually rendered uncompetitive?)  How are we to keep “AI sovereignty” for Americans?

If the top U.S. labs have to give up on what is supposed to be really important, they will look very “catchable.”  The policy creates a big incentive for some other parts of the world to invest a lot more in compute.  How safe is that?  And in the short run, do we also have to stop American firms from selling their compute abroad?

Does this whole thing mean that regular companies can use AI to write code, but the main labs cannot?  Solve for the equilibrium there.  Or the whole code writing function goes away for everybody?

What is the political mechanism through which RSI progress gets turned back on again, and when?  Is it as rational as our current debate over data centers?  If the top companies are no longer building RSI models, how are they supposed to know how safe or unsafe they might be?

At current margins, are humans even capable of writing the next steps of code that are required for further progress?  At whatever speed?

I do not find this to be a workable idea.  I think it is “people wanting to do something that sounds reasonable and less scary,” without thinking through the practical details.  I have made a related point before, but one does best with AI policy when you start with the stuff that otherwise you might put into your last two or three paragraphs of an essay (“It remains to be seen how…but we are all human beings and surely we should not give up hope…”).  Even if you believe in a deal with China, and I do not, we (and others) now know competitors can catch up more quickly than we had thought, at least if they think they have a chance of winning.  They might even buy some space compute from Elon.

C’mon people, I know for many of you this is a difficult take to swallow, because it feels like “we don’t care enough” if we do not take some dramatic steps.  Nonetheless the reality is that, a long time ago, we made a whole series of decisions that imply…we have no other choice than to just see this one through and get it more or less right.  The longer we deny that truth, the harder it will be for all of us.

The post Banning self-recursive improvement in AI models? appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/banning-self-recursive-improvement-in-ai-models.html/feed 94 93857 https://marginalrevolution.com/marginalrevolution/2026/09/further-results-on-ai-and-labor-market-reallocation.html?utm_source=rss&utm_medium=rss&utm_campaign=further-results-on-ai-and-labor-market-reallocation Further results on AI and labor market reallocation https://feeds.feedblitz.com/~/969438869/0/marginalrevolution~Further-results-on-AI-and-labor-market-reallocation.html https://marginalrevolution.com/marginalrevolution/2026/09/further-results-on-ai-and-labor-market-reallocation.html#comments Mon, 21 Sep 2026 18:49:42 +0000 https://marginalrevolution.com/?p=93851 This paper documents how artificial intelligence has affected stocks and flows in the U.S. labor market. Combining CPS and JOLTS data with AI exposure and adoption measures from OpenAI and Lightcast, respectively, we construct labor force stocks, worker flows, and market tightness by AI exposure and adoption. Since the introduction of LLMs, workers with high […]

The post Further results on AI and labor market reallocation appeared first on Marginal REVOLUTION.

]]>

This paper documents how artificial intelligence has affected stocks and flows in the U.S. labor market. Combining CPS and JOLTS data with AI exposure and adoption measures from OpenAI and Lightcast, respectively, we construct labor force stocks, worker flows, and market tightness by AI exposure and adoption. Since the introduction of LLMs, workers with high AI exposure and adoption have experienced larger declines in job-finding and job-switching rates than other groups, while their within-job activity switching has increased noticeably relative to others. The positive effects of AI on nominal wage growth and hours worked attenuate markedly in the post-LLM period, suggesting that LLM-driven reallocation has operated through within-firm task reorganization and a reconfiguration of labor inputs, alongside weaker demand for workers most exposed to AI. To quantify LLM-driven reallocation pressure, we construct individual-level hirability and separability indices and showthat dispersion in job-finding prospects has risen markedly since 2023, largely driven by the AI factors. The natural rate of unemployment—recovered from the trend components of unemployment inflows and outflows by AI exposure and adoption, as well as from a theoretical model of structural unemployment—is estimated to have risen by about 0.1-0.2 percentage point since the introduction of LLMs, albeit with considerable uncertainty.

That is from a new NBER conference paper by Hie Joo Ahn and Nicholas A. Carollo.  Via Inclusive Productivity Network and Alex Imas.

The post Further results on AI and labor market reallocation appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/further-results-on-ai-and-labor-market-reallocation.html/feed 13 93851 https://marginalrevolution.com/marginalrevolution/2026/09/monday-assorted-links-578.html?utm_source=rss&utm_medium=rss&utm_campaign=monday-assorted-links-578 Monday assorted links https://feeds.feedblitz.com/~/969436823/0/marginalrevolution~Monday-assorted-links.html https://marginalrevolution.com/marginalrevolution/2026/09/monday-assorted-links-578.html#comments Mon, 21 Sep 2026 17:57:03 +0000 https://marginalrevolution.com/?p=93853 1. Avi Loeb video on UAPs. 2. “[The Mises] library now has a permanent independent home: https://freecapitalists.org/. It holds 945 books, 3,374 journal issues and 2,732 recorded lectures, all free, in PDF, EPUB and MOBI, with the rights recorded per work.” 3. “They invited me to come see it for myself at Sopchoppy’s annual worm grunting festival, which […]

The post Monday assorted links appeared first on Marginal REVOLUTION.

]]> 1. Avi Loeb video on UAPs.

2. “[The Mises] library now has a permanent independent home: https://freecapitalists.org/. It holds 945 books, 3,374 journal issues and 2,732 recorded lectures, all free, in PDF, EPUB and MOBI, with the rights recorded per work.”

3. “They invited me to come see it for myself at Sopchoppy’s annual worm grunting festival, which draws thousands of visitors.” (NYT)

4. The (small) movement to return Europe to Latin.

5. Political pro-Trump bias in Polymarket?

The post Monday assorted links appeared first on Marginal REVOLUTION.

]]> https://marginalrevolution.com/marginalrevolution/2026/09/monday-assorted-links-578.html/feed 69 93853