huitzitziltzin

Of all the things we’re never going to do… we’re never going to agree with China to pause AI research in a verifiable way that would mean they could trust us or we could trust them. It’s just never going to happen.

With nuclear weapons you could conceivably go to a small list of specific places to inspect stocks of weapons. You can watch for tests from outer space or measure the quakes they create.

My read of the “deep seek moment” was that a lab came out of nowhere and trained a very powerful model with vastly less power than was previously required.

The AI labs are private, not owned or run by the government. They are distributed widely within and across countries. The barrier to entry is much lower than that for nuclear weapons.

It’s just a complete non-starter.

(And yeah I’m deliberately ignoring the obvious incentives Dario has to recommend this course of action as the owner of a leading lab himself.)

Example 6 is a good one. Iran attacked water infra in the US recently and maybe they would have done a “better” job (from their point of view) had they used Fable.

The “worst case” with 6 is potentially very bad but I think we are currently using advanced AI models to harden systems and patch vulnerabilities more aggressively than anyone is trying to bring down the whole power grid (for example).

I think it’s a potentially harmful case but my take is defensive capabilities are scaling as fast as offensive capabilities but defense is being implemented faster than anyone is going on offense?

Example 7 is Russia and Ukraine right now according to public information. It sounds like entirely autonomous weapons are deployed to the battlefield already. I put this in the “not likely to be a widespread problem” category for now.

I’m an economist. On the basis of current evidence, I view AI as a complement to human labor, not as a substitute for it. That’s the source of my rejection of the wild labor market disruptions theories.

I just don’t see any evidence yet that whole categories of jobs are being eliminated, with the single exception (so far!) of the end of “professional essay writing services for cheating college students,” and similar services.

That used to be a big business in Kenya, but is now effectively gone. (Covered in the New York Times this weekend if anyone is looking for the discussion.)

I did know about the mass shooting but failed to mention it here. I’d put it in the “unlikely to be a widespread problem” category. If we’re in the “one AI driven mass shooting every four years” world for example it’s fair to call it a rare issue.

The environmental impact seems either very overblown (e.g., water usage just isn’t that high) and the part that isn’t overblown is totally abatable (e.g., noise and emissions from gas generators). Nuclear or solar/renewables with batteries wouldn’t pollute.

I’ve seen no estimates of the additional deaths due to extra emissions specifically from power generation for AI purposes. If you have some, share them.

I’m willing to bet that they are a small rounding error against preventable deaths due to emissions from transport and non-AI-related power generation (which is an important and urgent issue worth spending a lot on, to be clear!). I’m happy to update that belief given evidence.

“ No other human activity poses this level of danger.”

I really, really disagree with that statement.

I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.

What’s the most dangerous thing that’s happened with an LLM so far? (This question is serious - maybe I don’t know the right examples.)

Example 1: I’m aware of a small number of people killing themselves in some kind of AI-facilitated psychosis. That is very unlikely to be a widespread problem.

Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.

Non-example 3: I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation. There’s no evidence for that.

Non-example 4: all the even-wilder Rationalist speculation about basilisks and the like is entirely divorced from reality.

I am looking for better reasons (supported by actual evidence!) to be more concerned than I am now: right now I am not concerned at all.

It’s kind of a fun story but it doesn’t make sense. Was that the only data in a data center in the world that’s not backed up elsewhere? Did I miss that part?

Newspapers have an extremely expensive product. They have to pay for it somehow! You can’t give away an expensive product for free forever!

No one on the internet likes paying for access to content. After 35 years we have not found a way to monetize except ad tech.

Is that so hard to understand?

Every time someone links an article on this website from an expensive print publication, there is immediately a link in the comments to a paywall-evading site!

The dialog around ads on HN is extremely low quality, highly focused on costs and with no attention at all paid to benefits.

I stand by my statement.

I’d love an estimate from you (or anyone) about the marginal effect on the profession’s “legitimacy” (which is what? and how’s it measured?) from having the prize include Nobel’s name vs. not including it.

Really we don’t care.

Sure…

Mokyr’s northwestern website has links to a lot of his papers.

An extremely crude selection rule:

Anything published in the American economic review, quarterly journal of economics, journal of political economy has the profession’s “highest stamp of approval”. It’s really hard to publish anything there. (There are two journals im not listing in that “top” category but he has no papers there on his website.). On aghion or howitts websites, look for the above journals but also econometrica and the review of economic studies. Those are the “top five” in the field.

There are surely papers in good history and Econ history journals on mokyr’s website but I don’t know the journals!

Standards for any chapter in a “handbook of X economics” or “handbook of the economics of X” are high - those should be good surveys.

Similarly a paper in the “annual review of economics”

Also mokyr has a bunch of work on Amazon. “The lever of riches” is a classic. “A culture of growth” is well regarded.

Finally he has a forthcoming book called “two paths to prosperity” with two other distinguished guys - one Econ historian (greif) and one political economy guy (tabellini). It’s coming out in about three weeks. Good timing, Princeton U Press!

Aghion and howitt have a growth textbook at the advanced undergrad level called “the economics of growth.”

They have a much more advanced work called “endogenous growth theory” which is for specialists (or at least anyone with first year PhD macro)

Aghion has a book called “the power of creative destruction.”

You can start with a used copy of an older edition of mankiw’s text which is a standard undergrad reference.

Barro also has an old undergrad macro text which is good

The gap between undergrad macro and professional macro is extremely large. That shouldn’t dissuade you it’s just a note.

Economist here…

An unexpected (to me!) prize but definitely a good one.

What’s notable is that mokyr’s research is very, very accessible to a layman. You can read his books and understand them nearly perfectly without needing substantial technical background. (Of course there’s a huge existing literature in economics and history he’s engaging with which you won’t know, but I’m not an economic historian either so a lot of it is unfamiliar to me too.). Try it! Hopefully you learn something.

Also the committee always releases a good non-technical summary of the laureates work and an even better “more technical” summary. You can start there for an overview.

As for the point which will be raised endlessly here that this is “not a real Nobel” - whatever. No one in the economics profession cares. Alfred Nobel doesn’t have a monopoly on prizes or priority to decide which fields are worth recognizing. It’s our highest prestige prize. Call it what you want.

If you read the review, you will see the author notes that issue pretty often and specifically discusses ways in which the oldest book he reviews (from 2014) both is and isn’t useful.

It’s worth reading

Why stand up and not use your legs at all?? People have rowed it.

This is a weird stunt that won’t prove anything. If he (magically) made it in a week people would still fly.

What’s the point? Don’t say “raising awareness”. Whose mind does the exercise have a chance of changing about what question? What behavioral change will that changed mind cause?

Agree that these are very very hard industries in which to secure employment.

At the least the anecdote is not that informative about the effects of AI, given the details.

(2021)

The funniest thing about the reaction from all of his neighbors is that none of them are willing to do the one thing that really would annoy Zuckerberg and probably make him move: build multi story, multi tenant housing (i.e., small apartment buildings!) on plots he doesn’t already own.

Go to the city and ask for permission to replace two lots with five story apartment buildings with a view into his backyard. He’ll sue immediately and if he can’t stop it he’ll move.

No doubt it’s impossible to even get permission to do so (which is a separate problem!), but if those whiners really want revenge on the guy who they think “ruined their neighborhood” there’s one option they aren’t pursuing.

Entirely useless recommendations and very dubious empirical claims here. Let’s take them one at a time….

- “The climate crisis and socio-ecological issues are broadly absent from economic curricula.”

-> I don’t believe that for a second. Externalities are taught in every Econ 101 class I’ve ever heard of.

“75% of universities do not teach any ecological economics”

-> as a whole class maybe not but that doesn’t mean the material doesn’t get covered.

“ instead, when issues of ecological sustainability are taught, environmental damage is considered as something that needs to be priced into market mechanisms.”

-> which is a completely normal, standard idea among economists for good reason!

“Economics education does not address historical and contemporary power imbalances”

-> that’s not our job? Wtf - that’s not part of economics at all. Does it get covered in statistics? In history? I don’t know who’s responsible but it’s not us.

“55% of universities do not provide meaningful teaching on questions of historical slavery, colonialism, or neocolonialism at all. History and ethics are absent from these discussions.”

-> economics departments are not being held responsible for this supposed omissions but I really doubt this supposed fact is true. Does an American history class count or not ? It’s just not possible that slavery is not taught.

“Mainstream neoclassical economics dominates the economic theories taught.”

-> also a good sign as that’s the standard. We don’t teach Marxist thought anymore. That’s progress.

“Of the 480 theory modules we graded, 88.3% of them included mainstream neoclassical economic thinking focusing on rational, self-interested individuals.”

-> Show me any alternative that’s credible. Also: behavioral methods are widely taught. We are plenty criticism of our own models. That’s also Econ 101 material.

“They are almost entirely taught through quantitative technical skills.”

->. Good. Also: as opposed to…? What exactly?

“Economics is taught in isolation from other social sciences. The discipline of economics should be embedded within the social sciences, and students should be encouraged to learn across other disciplines such as politics, sociology, geography, and history, but for the most part, it remains siloed”

-> that’s true of what happens in every other department too. Leave it up to universities to set distribution requirements.

“There are two programmes that are critical, climate-conscious, and provide an economics education fit for the 21st century. SOAS and the University of Greenwich introduce students to a range of intellectual and methodological perspectives within the economics discipline. They put a learning focus on climate, power, and inequality throughout the course.”

-> eye roll.

I have a PhD in economics with a specialization in antitrust policy.

Fact:

“Matt stoller gets big, important issues wrong in antitrust all the time and his analysis shouldn’t be taken seriously.”

Source: me

You are welcome to cite that.

Not surprising at all to hear that Matt Stoller got the economics of an antitrust issue wrong. I appreciate the attention he brings to the topic and his diligence in digging up stories but his own analysis is often just totally wrong. I don’t think he’s worth taking seriously as an analyst. Maybe as a reporter at best.

I would love to see more of their game theory example.

Having experience teaching the subject myself, what I saw on that page is about the first five minutes of the first class of the semester at best. The devil will very much be in the other 99% of what you do.

Great question! And a great example of why the central question in any antitrust issue is defining the market.

Define it narrowly enough, and Firm A buying Firm B might create a monopoly - the regulator cannot approve that.

Define it broadly, and Firm A buying Firm B takes the market from 21 firms to 20 - the regulator is unlikely to stand in the way.

(In practice the market definition question in the US will often happen in front of a judge and the two sides are the merging parties vs. the regulator (FTC or DOJ).)

In this case, is the relevant market “restaurant delivery service via an app” ?

In that case five firms is (likely) moderately concentrated at last. (If you have the market shares of the five firms, you can compute an index called the HHI to get a rough sense of how much more concentrated the market will be before and after a merger.)

However… “app based food delivery” doesn’t seem like a credible market to me. The firms are unlikely to have pricing power either over restaurants or customers.

Candidate competing markets:

- going to the restaurant directly

- making food at home

- delivery provided by the restaurant

Any one of the five firms currently in the market who wanted to buy another would argue to include all three as the “relevant market” in which case their share (both before and after) is tiny.

The regulator will argue for a narrower definition.

When you see commentary on this website about “X is a monopolist” or strong claims like that (which are made frequently!), almost never is there any appeal to a market definition!

Fortunately judges get better input than HN commentators.

(2018)

And the title should be “…who speak many languages.”

Speaking one language may be something of a mystery but to most of us it’s at least a familiar state.

How do you handle the very well known limits of LLMs in your especially-sensitive use case? Hallucinations are the leading example. Health queries are a really bad place to do even mild “imagining” of responses.

This feels like a “break glass in case of emergency” kind of moment. Sure there are no details yet, but I’m trying to imagine details which would make me think “that arrest makes sense.” If I were in Milwaukee I’d be in the streets.

Great piece and deserves to be complemented by the time series of daily outlays from the US treasury starting Jan 20 compared to the last two years.

Elon Musk is failing to cut American spending https://economist.com/finance-and-economics/2025/02/12/elon-... from The Economist

tl;dr: DOGE is having zero impact on federal spending.

The large tech firms get a surprisingly large amount of hate on antitrust issues on this website for startups so I appreciate your point bc I think it’s often missed.

The oil dividend (likely what they are talking about though I didn’t look) is something like $1,300/year. That’s not universal basic income in the pie-in-the-sky way the phrase is usually used. No one is interested in whether the Alaska dividend makes people not work.

If UBI is limited to a low amount that you couldn’t possibly live on then sure, it’s perfectly possible.

Of all the things which are never going to happen, I put universal basic income at the top of the list. Faster than light travel seems vastly more likely than universal basic income.

I don’t believe we can pay for it, nor do I believe that it’s remotely politically feasible.

Studies like this are perfectly nice but I can’t imagine they are actually designed to confidently predict the general equilibrium effects of rolling out UBI on an entire society. (The studies may not be designed with that in mind but it’s crucial.)

There are plenty of people working on the application of ML in economics, including NN methods, and including in macro.

That’s a totally different class of model to this agent based approach. People in the (small) agent based modeling community have been pushing their stuff for decades to no effect.

Sure it’s possible that there’s some amazing advance I can’t see coming in the future but as of now I would not recommend anyone pursue ABM.

I don’t agree that it doesn’t make sense. It’s a good approximation a lot of the time.

Also: from your comment I’m pretty sure you don’t have the background (correct me if that’s wrong) and so don’t know what it really means in practice to make an equilibrium assumption in a macro model: Markets clear, on average people have reasonable beliefs about the evolution of aggregate variables, firms maximize expected profits. That’s all pretty harmless.

I definitely don’t agree that we are “using mathematical tools which aren’t suited for that.” We aren’t treating the economy “like a steam engine.” The entire revolution in macro from the 1970’s on involves optimizing agents. There is no useful analogy to a steam engine.

I suspect you have been reading criticism by people who are misinformed about what macro actually is and how it is practiced.