Category: Current Affairs

China fact of the day

by  Tyler Cowen in

With surrogacy illegal in China, an industry of agencies, consultants and fertility clinics has emerged to connect clients with women overseas willing to carry their children. While there is no data on the number of children born to Chinese parents via surrogacy, a recent study showed nearly a third of intended parents for surrogate babies in the US were international and some 40 per cent of those were from China.

Early demand largely came from couples struggling with fertility. But there was growing interest from younger women physically capable of becoming pregnant but reluctant to accept its impact on their bodies and careers, said consultants, clients and doctors interviewed by the FT…

Other destinations include Georgia and Kyrgyzstan, where packages can cost as little as $63,000.

Here is more from Eleanor Olcott at the FT.

Words to live by?

by  Tyler Cowen in

In the past year American markets have digested the largest-ever initial public offering (SpaceX), equity raise by a public company (Google) or a foreign firm (SK Hynix), private-funding round (OpenAI) and private-debt deal (Broadcom), as well as most of the biggest bond issue in history (Amazon). There was the first $1trn exchange-traded fund, or ETF (Vanguard), a record-breaking cash pile (Berkshire Hathaway) and stock-buyback programme (Nvidia), and the consummation of the largest-ever leveraged buyout (Electronic Arts). Then there are the enormous mergers in railways (Union Pacific and Norfolk Southern), utilities (NextEra and Dominion Energy) and media (Paramount and Warner Bros Discovery)—the last of which led, this week, to the largest ever junk-bond offering, beating a record set only last week (SoftBank). New markets have been created out of thin air (for compute) or become much bigger (for predictions). Worries about massive corporate scandals (First Brands) and hedge-fund blow-ups (Situational Awareness), which would have once occupied investors’ attention for months, are steamrollered by the relentless, totalising and extraordinarily flexible machine that is American finance.”

An excellent passage, from The Economist.

A Normal Debate?

by  Alex Tabarrok in

Computer scientists Arvind Narayanan & Sayash Kapoor wrote AI as Normal Technology

The statement “AI is normal technology” is three things: a description of current AI, a prediction about the foreseeable future of AI, and a prescription about how we should treat it. We view AI as a tool that we can and should remain in control of, and we argue that this goal does not require drastic policy interventions or technical breakthroughs. We do not think that viewing AI as a humanlike intelligence is currently accurate or useful for understanding its societal impacts, nor is it likely to be in our vision of the future.

AI has responded in video format.

that led Kapoor to create his own diss track:

Both of these are good, great even, but the sneering in the first video is sublime. Keep in mind that this is almost entirely AI, the script, the music, the lyrics everything. It seems like just yesterday when people said AI could never be creative.

Doesn’t seem like a normal debate to me, regardless of which side you think won.

Alvin Roth to the rescue, the polity that is Singapore

by  Tyler Cowen in

Singapore has launched a dating platform, the latest social-engineering experiment by the city-state’s government to tackle its fast-declining fertility rate.

The initiative, known as FirstDate, opened under a pilot scheme this month for public sector employees aged 21-35 and uses a Nobel Economics Prize-winning matchmaking algorithm. An additional tool suggests date activities and allows users — who receive only one match at a time — to rate their experience in a survey.

The platform is the product of the annual hackathon held by the Singapore government’s technology agency earlier this year.

“FirstDate started with a question among a group of GovTech officers: does having more potential matches necessarily make it easier to find a suitable match?” the website said.

The app, which joins a crowded field of dating apps as well as more bespoke matchmaking services, is Singapore’s latest effort to reverse its falling birth rate, which has made the city-state one of the world’s fastest-ageing countries.

Here is more from Owen Walker at the FT.

Trump Administration Limits Predatory Lending in Education

by  Alex Tabarrok in

The New Republic writes “President Trump is banning students majoring in degrees that don’t make enough money from taking out college loans.” Yes, but do note that no student is banned from any major and the lending rule is mild. Undergraduate programs must show:

that their graduates earn more than the typical high school diploma holder…[and] graduate programs will be required to demonstrate that their graduates earn more than the typical bachelor’s degree holder. (emphasis added).

Think about how low that bar is. The comparison group for an undergraduate program is working adults aged 25-34 with nothing more than a high school diploma. A college program that can’t beat that has almost certainly made its students worse off. For graduate programs the bar is the lowest of several bachelor’s benchmarks, including bachelor’s holders in the same field. A master’s in social work need only beat people with a bachelor’s in social work. A program must also fail in two out of three years before it loses loan eligibility. The Department estimates that about 5% of programs will fail in the first year.

I mocked the term “predatory lending” when it first became common in the financial crisis but in this case predatory lending fits the bill because the real borrower isn’t the individual student. Under income-driven repayment, the taxpayer is a forced co-signer, and it’s the taxpayer who gets predated.

Most expansions of the student loan program have been motivated by the picture of an enterprising student who works hard and wants to major in mechanical engineering or nursing but because of their poor circumstances they can’t afford college. “Credit constraints, asymmetric information, you can’t collateralize human capital,” said the economists. Nice theory, what’s the practice?

The economists wanted loans for good investments and insurance against bad luck but the economists can’t swing the vote and once the government is lending, colleges want more tuition money and students want more forgiveness. The result is a subsidy for programs whose graduates are never likely to repay. As Looney and Yannelis document:

Starting in the late 1990s, policymakers weakened regulations that had constrained institutions from enrolling aid-dependent students. This led to rising enrollment of relatively disadvantaged students, but primarily at poor-performing, low-value institutions whose students systematically failed to complete a degree, struggled to repay their loans, defaulted at high rates, and foundered in the job market. As these new borrowers experienced similarly poor outcomes, their loans piled up, loan performance deteriorated, and with it the finances of the federal program.

Indeed, the program worked in reverse of what was promised. The biggest subsidies went to programs whose graduates were least able to repay, rather than programs with the strongest case for public support. As I wrote earlier:

Looney does a back of the envelope calculation and estimates that typical graduates in Mechanical Engineering will on average get a 0% subsidy but graduates in Music will get a 96% subsidy, in Drama a 99% subsidy and Masseuses a 100% subsidy on average. This of course is exactly the wrong approach. If we are going to subsidize, we should subsidize degrees with plausible positive spillovers not masseuses.

The courts later blocked Biden’s Save plan but the problem is built into income-driven repayment. If music, drama and masseuses are promised a 95%+ subsidy who is paying? The taxpayers. Moreover, it’s even worse than this because the very existence of these loans incentivizes the creation of expensive, useless programs. It’s not just the drama colleges, however. Not surprisingly, the law schools have proven adept at using Public Service Loan Forgiveness (PSLF) to rip off the taxpayer. The school raises tuition, then covers the student’s small income-driven payments for ten years, and the taxpayer forgives the rest. In short, protecting students from the cost of failure rewards colleges for producing it.

Fortunately, the same bill limiting loans ended Grad PLUS loans and capped graduate borrowing. You can see the logic: if taxpayers are going to insure the loans, they need some say over which programs qualify and how much is borrowed. I don’t like giving government that power, but this is the Mises–Higgs intervention ratchet in action: subsidize the loans, absorb the losses, then regulate the programs to limit the losses.

My ideal program would get the government out of the student loan business altogether but until then this is a good first step at limiting one of the most expensive and wasteful programs of the federal government.

Earth fact of the day, #2

by  Tyler Cowen in

The shortages have gone on for so long that they are aggressively driving down how much carbon is being released into the atmosphere, a Washington Post analysis of data from the International Energy Agency shows. People worldwide are using significantly less oil and gas, which means less climate pollution…

Such an annual decline has not happened since the height of the coronavirus pandemic. Fossil fuel consumption dropped significantly more then, and consumption was lower in absolute terms, too. Crude oil demand averaged 91 million barrels per day in 2020, compared with 102 million barrels per day under the latest IEA forecast.

But this year’s drop — especially given the sharp increase that was initially forecast — is substantial.

Here is the full story.  Not a good thing overall, but there is a lesson in that to…

Armenia notes

by  Tyler Cowen in

I strongly recommend a trip to Armenia, now is a great time to go.  Here are a few things I noticed:

1. The country remains intellectual.  The book fair is well attended.  The classical music season is excellent, with good concerts multiple times a week, and the theatre scene, in both Armenian and Russian, is active.  The country has only a little more than three million people, yet they have won gold medal in the chess Olympiad three times in recent memory.  You could say the country has maintained the (few) strengths of the former Soviet Union, and is arguably the only place that has.  It feels more like the former Soviet Union than Russia does.

2. The National Gallery has excellent Soviet and Russian paintings, while their Armenian collection is definitive.  The Medieval Manuscripts museum is first-rate.  Unfortunately, neither does anything to put images online, but all the more reason to go.

3. Armenian society is nearly 100% white and 100% Christian, as most of the immigrants (is that the right word here?) are Russian.  Crime is close to zero, and the cars stop when you try to cross the street.  Being a loyal northern Virginian, that mix is not “my thing.”  But if that is what you wish to experience, if only briefly, Armenia is the place that serves it up without apology.  Even secular Armenians have great respect for the national Apostolic church, as they view it as a guardian of Armenian independence.

4. The old monasteries are A level sights, and many of them are mere day trips from Yerevan, the capital.  Most of the tourists you see at these are Russians, not Westerners.  You also will see some Iranian tourists.

5. Yerevan is not great for old buildings, but the city is very walkable and pleasant and lively.  I believe the winters are brutal, but September there is wonderful, every day was perfect weather.

5b. In Armenia you will see how their brutalism in fact comes from their earlier religious architectural traditions, and how the newer churches embody modern brutalist ideas as well.  I found this very interesting.

6. I have a separate piece coming out on food in Armenia, for now I will just say it is excellent and novel as well.

7. Armenia is not dirt cheap, but it is much cheaper than most of Europe and so can be considered a bargain in relative terms.

8. Almost everyone in Yerevan speaks good Russian.  The influence of Russia is embedded in virtually every institution, people like Russia, Russia supplies most of their energy, and is the number one trading partner.  Russian troops protect them from Turkey.  A Russian does not even need a passport to enter and live in Armenia.  Russia is a permanent fixture in Armenian life, and most people do not seem to mind that.  They are very disappointed that Putin did not protect them more in the war with Azerbaijan, but that is their major source of discontent, not that Russia and Putin are evil.

9. Armenia has close relations with Iran, and Armenians do not seem to hate Iran.  They regard the country as a partial counterweight to Turkey, and for that matter Russia.  People in Armenia are very conscious of the x-risk represented by Turkey.  They also will speak to you about history in great detail, and they are very wrapped up in their past history.  They are very familiar with maps.

10. You can see Mount Ararat, supposed resting place of Noah’s Ark, from the outskirts of Yerevan.  It is beautiful.  It remains on the Armenian currency, and in the hearts of Armenians, yet it is now part of Turkey and is very likely to remain so.  Armenians refer to that part of Turkey as “Western Armenia.”

11. Here is someone else’s exaggerated take on the place, no it will not be the next Poland.

I would put it this way.  At any point in time, there are a few parts of the world that are perfect for visiting and viewing.  They should be comfortable, but still exotic and not overtouristed.  Not dangerous, and not too full of hassles, yet replete with challenging and stimulating experiences.  Yunnan in China is one such place.  Currently, aAzerbaijan, Georgia, and Armenia all fit this designation.  They all have complex histories, and are not always in great shape, so now is the time to go.  They will never get better to visit, and right now they are pretty close to perfect as destinations.  See what you can, while you can.

CA 91326: A Public Choice Mystery Investigation

by  Alex Tabarrok in

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.

My economics-rich Conversation with Gita Gopinath

by  Tyler Cowen in

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 Price of Intelligence is Falling Rapidly

by  Alex Tabarrok in

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.

Obama on agentic AI

by  Tyler Cowen in

Obama recently said:

“If we are thinking about AI just in terms of how do we cure cancer or get better energy, you can do that without having agentic AI and having it just roaming free in the internet.”

As Roon noted: “this sounds completely incoherent to me”

From Rob Saker:

An agent is a system that can plan, use tools, write code, query data, run experiments, check its own work, and keep going. That is how the work gets done. Treating “agentic” as optional decoration, as if the serious version of AI is a polite chatbot locked in a box while the unserious version “roams free on the internet”, is the kind of sentence you write when you’ve heard the buzzwords and never watched a lab actually use the technology. You do not cure cancer with a model that only answers questions. You cure cancer with systems that can read the literature, propose hypotheses, design assays, analyze results, rewrite the next experiment, and do that loop a thousand times faster than a human postdoc. That loop is agentic. Strip the agency out and you are left with a very expensive autocomplete.

I agree with those points, but my main concern is different.  How would we enforce such a prohibition on commercial agents?  Set up a Chinese-like firewall that bans Americans from accessing sites with foreign agents?  Monitor all those sites over time, so we know which suppliers to ban?  Ban VPN as well?  Give our government the power to inspect hard drives, in case agentic functions might be embedded there?  Set up FBI “phishes,” luring Americans in with the prospect of agent access from abroad, and then arresting them, as we do with child ****?

Something else?  How about insisting that all American (and foreign?) web sites set up tough captcha problems, so that agents may not be used (ha ha)?

You might also ask how an AI “agent” is to be defined, after all even O3 had some “agentic” abilities, such as opening up museum web sites to see which exhibits are on.  Is it typing things into boxes and filling out forms that is to be prohibited?  How much regulation of software would that require, how would that regulation actually occur, and what else would end up being restricted?

I do not think Obama intends to be supporting massive restrictions on freedom of speech and civil liberties.  Rather this is a good example of how some political factions will simply have “ideas about what might be good,” without having freedom concerns — or for that matter practicality and civil liberties concerns — center of mind in the first place.

And this would cause the immediate bankruptcy of both Anthropic and OpenAI, right?

I can readily imagine that, upon the advent of agentic AI, we need some significant changes in our cyber laws.  Now would be a good time to both commission and also cite some peer-reviewed academic literature on this matter!?  That is not an impossible thing to do, yet our public discourse seems oddly resistant to the notion.

Addendum: And do not forget the classic Gwern piece, remarkably early in its prescience.

UK productivity higher than we had thought

by  Tyler Cowen in

UK productivity rose at almost twice the pace previously estimated in the decade after the 2008 global financial crisis, according to new official figures that make Britain look much less of an international outlier.

Output per hour worked, the main measure of labour productivity, grew by an average of 1.3 per cent per year between 2009 and 2019, the Office for National Statistics said on Thursday. The ONS had previously estimated average annual growth of 0.7 per cent over the period.

The figure is based on a new “component-based” methodology that has already been adopted by many other national statistics institutes. It draws on a range of data sources to improve the ONS’s count of both the number of workers contributing to UK output and the number of hours worked.

Here is more from Delphine Strauss at the FT.

The Age of Wonders and Terrors

by  Alex Tabarrok in

Scott Aaranson, theoretical computer scientist writes:

My position on AI is merely the conservative, skeptical position of 2006, updated with intellectual honesty for the reality of late 2026. And that position, if you need me to spell it out, is as follows:

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

It seems to me that the Singularity has already started; it’s just wildly unevenly distributed. Yes, I still unload the dishwasher and clip my toenails. On the other hand, in whatever years I have left, I don’t expect that I’ll ever again prove a theorem because I’m actually needed to prove it. If I do, it will only be for my or others’ enjoyment or edification.

The test is this: if we took the news of these past few weeks and sent it back in time twenty years, would I agree that it looked like the beginning of an AI Singularity? The intellectually honest answer is: yes, absolutely. But then that’s all we need. No backsies.

I feel like it would be healthy for everyone to stop grinding their ideological axes, their sentiments about Dario Amodei or Sam Altman, for long enough simply to acknowledge that the wonders and terrors are here. They couldn’t be here more clearly if the sky had turned reddish-orange like in the Matrix movies.

It’s here clearly enough that, when I put my kids to sleep at night, I now feel it in the pit of my stomach: what sort of future can they possibly have? What could they learn today that could possibly be relevant to that future? (Yesterday, my 13-year-old daughter joked unprompted that, if she wants to become a mathematician, it now looks like she has maybe two more weeks.) Certainly when my grad students want to discuss what sort of careers might await them on graduation, I no longer have any clue what to tell them.

…For anyone who says AI doom sounds like an apocalyptic religion, that the rationalists/Singulatarians seem like a Bay Area cult, that Eliezer Yudkowsky gives off the vibes of a messianic prophet: yes, yes, and yes. But crucially, today you’re no longer being asked to believe in arguments and extrapolations, but only in the front-page news. Accepting the reality of the coming machine god after it’s solved Navier-Stokes and dozens of other longstanding open math problems (while dramatically ramping up in capability every month), is sort of like accepting Jesus after he’s returned to earth on the gleaming cloud. It’s the epistemic bare minimum.

Yes, there’s still enormous uncertainty about what the rest of our lives will look like, but as far as I can tell, there’s no longer any real uncertainty that it’ll all mostly revolve around AI, and the extent to which we succeed or fail at directing its power toward human flourishing.

By any accounting that doesn’t stack the deck, Eliezer Yudkowsky was right about what the greatest challenge facing civilization in our lifetimes was going to be, and you and I were wrong about it. Why I was wrong is a question I’ll ask myself every day in whatever time remains. But, you know, at least I updated once the prophesied wonders and terrors actually started arriving! If you haven’t done likewise, why haven’t you?

METR, EA, and others being attacked

by  Tyler Cowen in

Always focus on what you can learn from people and groups.  Criticize in ways where you can learn something from the answer (or non-answer) and from the dialogue.  If you attack how people look, their sex lives, their donors, whatever, it will make the critic stupider.  It may not even hurt the target of the attack, as it provides valuable publicity and also makes them look powerful.  Furthermore, there is nothing per se wrong with being “weird.”  What does that really mean anyway?  I’ve spent much of my life looking for “the weird,” though I would not frame it in those terms.