Category: Web/Tech

Surveillance is growing

by  Tyler Cowen in

California residents who launched fireworks for the 4th of July have tickets coming in the mail, thanks to police drones that were taking note. One resident, for example, racked up $100,000 in fines last summer due to the illegal use of fireworks. “If you think you got away with it, you probably didn’t,” said Sacramento Fire Department Captain Justin Sylvia. “What may have been a $1,000 fine for one occurrence last year could now be $30,000 because you lit off so many.” Homeowners who weren’t even present at the property also have tickets coming in the mail due to the social host ordinance.

Here is the source.  Elsewhere (NYT):

Hertz and other agencies are increasingly relying on scanners that use high-res imaging and A.I. to flag even tiny blemishes, and customers aren’t happy…

Developed by a company called UVeye, the scanning system works by capturing thousands of high-resolution images from all angles as a vehicle passes through a rental lot’s gates at pickup and return. A.I. then compares those images and flags any discrepancies.

The system automatically creates and sends damage reports, Ms. Spencer said. An employee reviews the report only if a customer flags an issue after receiving the bill. She added that fewer than 3 percent of vehicles scanned by the A.I. system show any billable damage.

I await the next installment in this series.

Grok 4 on economics

by  Tyler Cowen in

My prompt:

What is the best analysis of the incidence of the corporate income tax? How much falls on capital, labor, and the consumer, respectively? In the U.S. What does it work out that way?

Here is the answer, plus my response and its follow-up.  For one thing, it is the existence of the non-corporate sector, where capital may be allocated, that is key to getting off on the right foot on this question…

Emotions and Policy Views

by  Tyler Cowen in

I would call this a story of negative emotional contagion:

This paper investigates the growing role of emotions in shaping policy views. Analyzing social citizens’ media postings and political party messaging over a large variety of policy issues from 2013 to 2024, we document a sharp rise in negative emotions, particularly anger. Content generating anger drives significantly more engagement. We then conduct two nationwide online experiments in the U.S, exposing participants to video treatments that induce positive or negative emotions to measure their causal effects on policy views. The results show that negative emotions increase support for protectionism, restrictive immigration policies, redistribution, and climate policies but do not reinforce populist attitudes. In contrast, positive emotions have little effect on policy preferences but reduce populist inclinations. Finally, distinguishing between fear and anger, we find that anger exerts a much stronger influence on citizens’ policy views, in line with its growing presence in the political rhetoric.

That is from a new paper by Eva Davoine, Stefanie Stantcheva, Thomas Renault, and Yann Algan.

The Impact of Dating Apps on Young Adults: Evidence From Tinder

by  Tyler Cowen in

Online dating apps have transformed the dating market, yet their broader effects remain unclear. We study Tinder’s impact on college students using its initial marketing focus on Greek organizations for identification. We show that the full-scale launch of Tinder led to a sharp, persistent increase in sexual activity, but with little corresponding impact on the formation of long-term relationships or relationship quality. Dating outcome inequality, especially among men, rose, alongside rates of sexual assault and STDs. However, despite these changes, Tinder’s introduction did not worsen students’ mental health, on average, and may have even led to improvements for female students.

That is from a new paper by Berkeren Büyükeren, Alexey Makarin, and Heyu Xiong.

A consumption basket approach to measuring AI progress

by  Tyler Cowen in

Many AI evaluations go out of their way to find hard problems.  That makes sense because you can track progress over time, and furthermore many of the world’s important problems are hard problems, such as building out advances in the biosciences.  One common approach, for instance, is to track the performance of current AI models on say International Math Olympiad problems.

I am all for those efforts, and I do not wish to cut back on them.

Still, they introduce biases in our estimates of progress. Many of those measures show that the AIs still are not solving most of the core problems, and sometimes they are not coming close.

In contrast, actual human users typically deploy AIs to help them with relatively easy problems.  They use AIs for (standard) legal advice, to help with the homework, to plot travel plans, to help modify a recipe, as a therapist or advisor, and so on.  You could say that is the actual consumption basket for LLM use, circa 2025.

It would be interesting to chart the rate of LLM progress, weighted by how people actually use them.  The simplest form of weighting would be “time spent with the LLM,” though probably a better form of weighting would be “willingness to pay for each LLM use.”

I strongly suspect we would find the following:

1. Progress over the last few years has been staggeringly high, much higher than is measured by many of the other evaluations  For everyday practical uses, current models are much better and more reliable and more versatile than what we had in late 2022, regardless of their defects in Math Olympiad problems.

2. Future progress will be much lower than expected.  A lot of the answers are so good already that they just can’t get that much better, or they will do so at a slow pace.  (If you do not think this is true now, it will be true very soon.  But in fact it is true now for the best models.)  For instance, once a correct answer has been generated, legal advice cannot improve very much, no matter how potent the LLM.

As in standard economics, consumption baskets change over time, and that can lead to different measures of progress (or in the economics context, different estimates of advances in living standards, depending on whether the ex ante or ex post bundle weights are used).  Researchers could attempt the more speculative endeavor of estimating how LLMs will be used five years from now in everyday life (which will differ from the status quo), and then track progress on that metric, using those value weights.  “How rapidly are we improving these systems on their future uses?”

This alternate consumption basket approach gives you a very different perspective on progress in AI.

Note also that the difference between the “Math Olympiad measurements of AI progress” and the “consumption basket measurements of AI progress” may iincrease over time, especiallly if the basket of everyday uses does not change radically.  The everyday uses will peak out near maximum levels of performance, but there will always be a new series of very hard problems to stump the AIs.  It will become increasingly unclear exactly how much AI progress we really are making.

Balaji on AI

by  Tyler Cowen in

A few miscellaneous thoughts.

(1) First, the new bottleneck on AI is prompting and verifying. Since AI does tasks middle-to-middle, not end-to-end. So business spend migrates towards the edges of prompting and verifying, even as AI speeds up the middle.

(2) Second, AI really means amplified intelligence, not agentic intelligence. The smarter you are, the smarter the AI is. Better writers are better prompters.

(3) Third, AI doesn’t really take your job, it allows you to do any job. Because it allows you to be a passable UX designer, a decent SFX animator, and so on. But it doesn’t necessarily mean you can do that job *well*, as a specialist is often needed for polish.

(4) Fourth, AI doesn’t take your job, it takes the job of the previous AI. For example: Midjourney took Stable Diffusion’s job. GPT-4 took GPT-3’s job. Once you have a slot in your workflow for AI image gen, AI code gen, or the like, you just allocate that spend to the latest model.

(5) Fifth, killer AI is already here — and it’s called drones. And every country is pursuing it. So it’s not the image generators and chatbots one needs to worry about.

(6) Sixth, decentralized AI is already here and it’s essentially polytheistic AI (many strong models) rather than monotheistic AI (a single all-powerful model). That means balance of power between human/AI fusions rather than a single dominant AI that will turn us all into paperclips/pillars of salt.

(7) Seventh, AI is probabilistic while crypto is deterministic. So crypto can constrain AI. For example, AI can break captchas, but it can’t fake onchain balances. And it can solve some equations, but not cryptographic equations. Thus, crypto is roughly what AI can’t do.

(8) Eighth, I think AI on the whole right now is having a decentralizing effect, because there is so much more a small team can do with the right tooling, and because so many high quality open source models are coming.

All this could change if self-prompting, self-verifying, and self-replicating AI in the physical world really gets going. But there are open research questions between here and there.

Here is the link to the tweet.

The objectivity of Community Notes?

by  Tyler Cowen in

We use crowd-sourced assessments from X’s Community Notes program to examine whether there are partisan differences in the sharing of misleading information. Unlike previous studies, misleadingness here is determined by agreement across a diverse community of platform users, rather than by fact-checkers. We find that 2.3 times more posts by Republicans are flagged as misleading compared to posts by Democrats. These results are not base rate artifacts, as we find no meaningful overrepresentation of Republicans among X users. Our findings provide strong evidence of a partisan asymmetry in misinformation sharing which cannot be attributed to political bias on the part of raters, and indicate that Republicans will be sanctioned more than Democrats even if platforms transition from professional fact-checking to Community Notes.

Here is the full paper.  I guess it agrees with Richard Hanania…

My excellent Conversation with Austan Goolsbee

by  Tyler Cowen in

Here is the audio, video, and transcript.  Here is part of the episode summary:

A longtime professor at the University of Chicago’s Booth School and former chair of the Council of Economic Advisers under President Obama, Goolsbee now brings that intellectual discipline—and a healthy dose of humor—to his role as president of the Federal Reserve Bank of Chicago.

Tyler and Austan explore what theoretical frameworks Goolsbee uses for understanding inflation, why he’s skeptical of monetary policy rules, whether post-pandemic inflation was mostly from the demand or supply side, the proliferation of stablecoins and shadow banking, housing prices and construction productivity, how microeconomic principles apply to managing a regional Fed bank, whether the structure of the Federal Reserve system should change, AI’s role in banking supervision and economic forecasting, stablecoins and CBDCs, AI’s productivity potential over the coming decades, his secret to beating Ted Cruz in college debates, and more.

Excerpt:

COWEN: Okay, if the instability comes from the velocity side, that means that we should favor a monetary-growth rule to target the growth path of a nominal GDP, M times V, right?

GOOLSBEE: [laughs] Yes, and now you’re going to get me in trouble, Tyler. Here’s the thing I’ve known —

COWEN: You can just say yes. You’re not in trouble with me.

GOOLSBEE: I’m not going to say yes because, remember, I don’t like making policy off accounting identities. There’s no economic content in accounting identity. If you are trying to design a rule, that rule may work if the shocks are the same as what they always were in previous business cycles. I called it the golden path.

When we came into 2023, you’ll recall the Bloomberg economists said there was a 100 percent chance of recession in 2023. They announced it at the end of 2022. That’s when I came into the Fed system, the beginning of ’23.

That argument was rooted in the past. There had never been a drop of inflation of a significant degree without a very serious recession. Yet in 2023, there was. Inflation fell almost as much as it ever fell in one year without a recession. If you over-index too much on a rule that implicitly is premised on that everything is driven by demand shocks, I just think you want to be careful over-committing.

COWEN: I’m a little confused at the theoretical level. On one hand, you’re saying M times V is an identity, but on the other hand, it drives inflation dynamics.

GOOLSBEE: It’s why I started back from the . . . I bring a micro sentiment to the thinking about causality and supply and demand. I sense that you want to bring us to a, let’s agree on a monetary policy rule, and I’m inherently a little uncomfortable. I want to see what the rules say, but I fundamentally don’t want us to pre-commit to any given rule in a way that’s not robust to shocks.

COWEN: Now, you mentioned the post-pandemic inflation and the role of the supply side. When I look at that inflation, I see prices really haven’t come back down. They’ve stayed up, and I see service prices are also quite high and went up a lot, so I tend to think it was mostly demand side. Now, why is that wrong?

GOOLSBEE: There’re two parts to that. I won’t say why it’s wrong, but here are my questions. If you’re firmly a ‘this-all-came-from-demand’ guy, (A) you’ve got to answer, why did inflation begin soaring in the US when the unemployment rate is over 6 percent? Or we could turn it into potential output terms if you want, but output is below our estimate of potential. Unemployment is way higher than what we think of as the natural rate, and inflation is soaring. That already should make you a little questioning.

COWEN: I can cite M2. You may not like it. M2 went up 40 percent over a few-year period, right?

GOOLSBEE: Two, the fact that the inflation is taking place simultaneously in a bunch of countries of similar magnitudes that did not have the kind of aggregate demand, fiscal or monetary stimulus that we had in the US is also a little bit of a puzzle.

Then the third is, if you don’t think it was supply, then you need to have an explanation for why, when the stimulus rolls off, everything about the stimulus is delta from last year. We pass a big fiscal stimulus, we have substantial monetary stimulus that rolls off, the inflation doesn’t come down. Then in ’23, when the supply chain begins to heal, you see inflation come down. Those three things suggest there’s a little bit of a puzzle if you think it was all demand.

COWEN: No, I don’t think it was all demand, but you mentioned other countries. Switzerland and Japan — they import a lot. They were more restrained on the demand side. They had much lower rates of price inflation. That seems to me strong evidence for being more demand than supply.

GOOLSBEE: Wait a minute.

COWEN: I’m waiting.

GOOLSBEE: You’re going to bring in Japan?

COWEN: Yes.

GOOLSBEE: And you’re going to try to claim that Japan’s low inflation is the result of something in COVID? Japan had lower inflation all along, for decades before. They were going through deflation.

COWEN: But if it was mostly supply, a supply shock would’ve gotten them out of the earlier deflation, right? A demand shock would not have.

Recommended.

One possible reason why the skill premium is declining

by  Tyler Cowen in

This is especially true for those jobs that require the rudimentary use of technology. Until relatively recently, many people could get to grips with a computer only by attending a university. Now everyone has a smartphone, meaning non-graduates are adept with tech, too. The consequences are clear. In almost every sector of the economy, educational requirements are becoming less strenuous, according to Indeed, a jobs website. America’s professional-and-business services industry employs more people without a university education than it did 15 years ago, even though there are fewer such people around.

Here is more from The Economist, quite a good piece.  Of course this is also a reason why smart phones are underrated.

Are cultural products getting longer?

by  Tyler Cowen in

Ted Gioia argues that cultural products are getting longer:

Some video creators have already figured this out. That’s why the number of videos longer than 20 minutes uploaded on YouTube grew from 1.3 million to 8.5 million in just two years…

Songs are also getting longer. The top ten hits on Billboard actually increased twenty seconds in duration last year. Five top ten hits ran for more than five minutes…

I’ve charted the duration of [Taylor] Swift’s studio albums over the last two decades, and it tells the same story. She has gradually learned that her audience prefers longer musical experiences…

I  calculated the average length of the current fiction bestsellers, and they are longer than in any of the previous measurement periods.

Movies are getting longer too.  Of course this is the exact opposite of what the “smart phones are ruining our brains” theorists have been telling us.  I think I would sooner say that the variance of our attention spans is going up?  In any case, here is part of Ted’s theory:

  1. The dopamine boosts from endlessly scrolling short videos eventually produce anhedonia—the complete absence of enjoyment in an experience supposedly pursued for pleasure. (I write about that here.) So even addicts grow dissatisfied with their addiction.
  2. More and more people are now rebelling against these manipulative digital interfaces. A sizable portion of the population simply refuses to become addicts. This has always been true with booze and drugs, and it’s now true with digital entertainment.
  3. Short form clickbait gets digested easily, and spreads quickly. But this doesn’t generate longterm loyalty. Short form is like a meme—spreading easily and then disappearing. Whereas long immersive experiences reach deeper into the hearts and souls of the audience. This creates a much stronger bond than any 15-second video or melody will ever match.

An important piece and useful corrective.

Does AI make us stupider?

by  Tyler Cowen in

That is the topic of my latest Free Press column, responding to a recent study out of MIT.  Here is one excerpt:

To see how lopsided their approach is, consider a simple parable. It took me a lot of “cognitive load”—a key measure used in their paper—to memorize all those state capitals in grade school, but I am not convinced it made me smarter or even significantly better informed. I would rather have spent the time reading an intelligent book or solving a math puzzle. Yet those memorizations, according to the standards of this new MIT paper, would qualify as an effective form of cognitive engagement. After all, they probably would have set those electroencephalograms (EEGs)—a test that measures electrical activity in the brain, and a major standard for effective cognition used in the paper—a-buzzin’.

The important concept here is one of comparative advantage, namely, doing what one does best or enjoys the most. Most forms of information technology, including LLMs, allow us to reallocate our mental energies as we prefer. If you use an LLM to diagnose the health of your dog (as my wife and I have done), that frees up time to ponder work and other family matters more productively. It saved us a trip to the vet. Similarly, I look forward to an LLM that does my taxes for me, as it would allow me to do more podcasting.

If you look only at the mental energy saved through LLM use, in the context of an artificially generated and controlled experiment, it will seem we are thinking less and becoming mentally lazy. And that is what the MIT experiment did, because if you are getting some things done more easily your cognitive load is likely to go down.

But you also have to consider, in a real-world context, what we do with all that liberated time and mental energy. This experiment did not even try to measure the mental energy the subjects could redeploy elsewhere; for instance, the time savings they would reap in real-life situations by using LLMs. No wonder they ended up looking like such slackers.

Here is the original study.  Here is another good critique of the study.