In prior installments in this series of posts I have reflected on the Industrial Revolution and the Digital Revolution which were alike in many ways but differed in some significant ways as well. In this installment I will consider the ongoing Artificial Intelligence Revolution. As we will see, it is different in some basic ways from its predecessors.

The Industrial Revolution mechanized human muscle. The Digital Revolution mechanized human memory and communication. The AI Revolution mechanizes human judgment itself.

A good starting point is this post by Emily Chamlee-Wright at Persuasion:

Given the nature of my work, I’m in coalitions—and a lot of conversations—focused on fortifying American democracy. And like everyone, the promises and threats associated with the rapid advancement of AI technology are top of mind for me.

When it comes to scientific discovery, optimism abounds. The promise that AI tools like AlphaFold will accelerate biomedical discovery—from new antibiotics to deeper insights into the proteins driving Alzheimer’s and cancer—sparks excitement. But there’s also a pessimism that lingers in these conversations, especially when it comes to what the future holds for workers who are not at the frontiers of science. If you write code or oversee routine managerial processes, the thinking goes, your days of productive employment are numbered. And a jobless citizenry does not bode well for democracy.

Dr. Chamlee-Wright’s central argument is the human imagination will always come up with new “needs” and that every newly perceived human need will continue to clear through human labor markets. She assumes, without either theoretical or empirical support, that newly imagined human needs will remain monetizable in a labor sense once cognitive scarcity disappears.

Dr. Chamlee-Wright is ignoring a crucial difference between the AI Revolution and the Industrial and Digital Revolutions before it. Cognition not being scarce has never occurred before. There is no historical analogy, because in every prior revolution, the scarce factor remained human. This time, it does not. Not only is there no empirical evidence from current AI diffusion that points in that direction, in fact the early evidence points the other way.

Consider this example. The Indian Ministry of Statistics and Program Implementation produces a report called the “Periodic Labour Force Survey” (PLFS) annually. Here are the results for low-level IT support workers for the last five years:

In human terms that reflect a decline in 2019-2020, a post-pandemic spike, a steep subsequent drop, and continuing slow decline. One could argue that this volatility merely reflects pandemic distortion. But that interpretation fails to explain the lack of rebound once demand normalized. That is precisely the pattern seen in prior waves of clerical automation.

Not only does the 2021-2022 spike not refute the notion of the erosion of jobs by AI, it actually confirms it. That was the same phenomenon that was observed among secretaries, typists, and travel agents before their jobs were automated. Organizations routinely overhire during system transitions to handle data cleaning, exception processing, and customer education. Those roles disappear once automation stabilizes. This is not cyclical recovery; it is terminal transition.

The 2022-2023 spike reflects the adoption of self-service portals, AI-driven chat, and internal copilots. As the graph reflects that process is continuing. There is no cyclic recovery. Furthermore, due to factors characteristic of the Indian economy including labor surplus, wage compression, and strong substitution effects, in India headcount may remain stable while task content is hollowed out which is exactly what my analysis has suggested. If even in a labor-surplus, low-wage economy like India we see hollowing out, the effect in high-wage Western economies will be more severe, not less. India is not an outlier here; it is a leading indicator.

India is not an arbitrary example. I chose it precisely because it is a labor-surplus, low-wage economy with relatively credible national statistics and deep integration into global IT services. If cognitive automation were merely a rich-country phenomenon, India should be resilient. If anything, it should absorb displaced work. The fact that we instead observe hollowing out even here makes the pattern more, not less, concerning. It’s not cherry-picking. I chose it before I knew the results, confident in what it would demonstrate which it did.

Contrary to Dr. Chamlee-Wright’s expressed view of the resilience of democracy despite job loss, that reflects a misunderstanding of the challenge. The problem is not that democracy will survive despite job loss. The problem is that democracy was architected around a laboring citizenry. Remove labor, and you remove one of its structural pillars.

The Challenge of Artificial Intelligence: Part I—the Industrial Revolution
The Challenge of Artificial Intelligence: Part II—the Digital Revolution

I recently read the credible claim that the most viral posts on Facebook were now AI-generated “slop”. To my eye that has some serious implications for Facebook, the World Wide Web, AI, the world, and human cognition itself.

For Facebook, when the dominant content in a major social space is no longer human expression amplified by machines, but machine expression optimized for machines that drastically diminishes its value proposition.

The World Wide Web in its original form consisted of content generated by humans, links, and discoverability. A multi-billion dollar business, Google (now Alphabet) was built on that. Google and its emerging reliance on AI, replacing links with unattributed or pseudo-attributed AI-generated content, is already subverting that. Social media’s replacement of links with feeds had already eroded that. Now the replacement of human-created posts with AI-generated content i.e. “slop” further dilutes the value proposition. Feeds killed navigation; AI is killing authorship.

AI content generation has many advantages for the companies. It is cheap, scalable, easily tested in realtime, and can be quickly tailored for different demographics, moods, and trends. It is not equally advantageous for those who are notionally its customers.

In my earlier discussion of the Digital Revolution I held back from exploring a specific aspect of that revolution that is relevant here. One of the results of the Digital Revolution was the emergence of a large, globally distributed pool of workers who are digitally literate, technically competent, capable of speaking English content creators. Large segments of these roles as transcribers, customer support staff, content moderators, search engine optimizers, and coders are being automated or commoditized. Needing alternative sources of revenue they are turning to prompt farming, clickbait generation, fake personae synthetic virality, and scam-adjacent content to replace that income. It is the application of 19th century urbanization to the 21st century Internet.

Different from Marx’s predictions they are not politically dangerous. They are civilizationally dangerous.

In previous eras writing, printing, and distribution were all expensive. Now with a laptop, electricity, and an Internet connection someone with the knowledge to do it can create practically infinite output. That reduces the marginal cost of persuasion to something approaching zero. Where persuasion once required institutions, reputations, and capital, it now requires only throughput and targeting. The net result is a global factory for cognitive junk food.

These workers are not producing historically grounded argument, culturally embedded nuance, or philosophically coherent positions. They are producing memes, emotional triggers, visual hooks, and slogans.

I have also previously mentioned the weak Sapir-Whorf hypothesis. It is a robust finding among knowledgeable scholars, confirmed by empirical evidence. The cognitive effects of the flood of “slop” with which people are being inundated will be considerable. These are the effects of increasing “visualcy” including a diminishing in abstract reasoning, increasingly emotional expression, and reduced tolerance for ambiguity among others.

This is the cycle. AI displaces workers; displace workers produce AI-assisted junk; platforms reward it; AI trains on it; cognition degrades; demand for nuance collapses; displacement accelerates.

I just read a very encouraging post about the protests in Iran by Iranian journalist Roohola Ramezani at Persuasion. Here’s a telling snippet, its conclusion:

Iranians want neither a better version of the Islamic Republic nor a utopian revolution. They want the extraordinary era of Iranian history to end. They want to return to being a normal nation state that prioritizes its borders over ideological “frontiers,” its citizens over “martyrs,” and its future over seventh-century grievances.

A fierce, secular realism is emerging. The protest movement is redefining Iranian identity not through the lens of a “Global South” struggle against the West, but as a struggle to rejoin the West’s political and economic orbit. It’s a movement that does not fit neatly into the categories of anti-imperialism or identity politics; but it is, nevertheless, perhaps the most authentic democratic project of our time.

My problem with it is that I don’t see an evidentiary basis for that level of specificity. Quite to the contrary another alternative is that Iranians are reacting to the incompetence of the mullahocracy. What Iranians want in aggregate or whether they have common goals other than escape from their present misery I have no idea.

I have little doubt that some Iranians want precisely what the author suggests. Whether that’s all Iranians, a majority, or just a few I have no idea and I see no indication that anyone has the data required to make that determination. Shared opposition to a regime does not, by itself, imply shared agreement on what should replace it. The historical record suggests that one faction will succeed in applying its own preferred solution and that faction will not necessarily be the most numerous or even the most popular one.

Here’s my question. How would we arrive at an empirically-based understanding of what Iranians want right now? That’s not a rhetorical question. I genuinely want to know because I think that policies should be based on facts not wishful thinking, however heartfelt.

I’ll trust claims about national political aims only when they’re supported by (1) survey evidence with methods described, plus (2) at least one independent corroborating stream. That could be behavioral indicators, slogan analysis, turnout or legitimacy signals, or multiple polling efforts. Otherwise, I will treat them as hopeful interpretation.

In an op-ed in the New York Times Elizabeth Dickinson describes the situation on Venezuela’s border. The Chavista Maduro regime had delegated control of the border to the Colombian National Liberation Army (ELN), a guerrilla group founded to mount a leftist insurgency. Here’s the meat of the op-ed:

Now the E.L.N. stands emboldened to challenge the authority of the Colombian state — and U.S. ambitions in Venezuela. The borderlands are webbed with lucrative corridors where the E.L.N. and other armed groups move seamlessly and often exercise more control than the government. With profits flowing from illegal mining, drug trafficking and human smuggling, both the Colombian guerrillas and complicit members of Venezuela’s security forces have deep interests in maintaining the status quo in Caracas and resisting attempts to bring rule of law to these territories.

In advance of Mr. Maduro’s capture, the E.L.N. was taking steps to ensure its interests in the borderlands were safe, regardless of what happened in Caracas. Since mid-December, it has gone on the offensive in the Colombian region of Catatumbo, displacing thousands of civilians in the process. It has also clashed with a local criminal group known as the 33rd Front, a dissident faction of the Revolutionary Armed Forces of Colombia, or FARC, which has repeatedly angered the E.L.N. with attempts to control key rivers used for trafficking in and out of Venezuela. President Gustavo Petro’s announced deployment of some 30,000 troops to the border has done little to stop the fighting.

But rather than anchoring the region with America’s longtime partner in Bogotá, President Trump turned on Mr. Petro, threatening direct attacks on Colombia the day after Mr. Maduro’s capture on Jan. 3. Although a phone call last Wednesday between the leaders lowered tensions, the détente is fragile.

Removing (literally) Maduro from Venezuela may have the unforeseen effect of destabilizing not just Venezuela but Colombia as well.

Aristotle said it more than two millennia ago: nature abhors a vacuum. Already, clashes between armed groups have forced thousands of civilians from their homes, and a vacuum on the Venezuelan side could intensify competition among criminal and insurgent networks. If the Chavista regime, having never established effective control of the border, does not stabilize the situation, and the United States which lacks both capacity and strategy does not either, the ELN, FARC dissidents, cartels, paramilitaries, or others will. The U.S. lacks a coherent stabilization strategy, and its recent actions risk exacerbating instability unless paired with a credible regional security plan. Failing that neither the Venezuelan people, the Colombian people, or we may like the result.

At Reuters David Lawder reports:

WASHINGTON, Jan 13 (Reuters) – The U.S. government posted a $145 billion budget deficit for December, up 67% or $58 billion from a year earlier due to record outlays that were inflated by calendar shifts in benefit payments and receipts, the Treasury Department said on Tuesday.

The report showed that revenue growth from President Donald Trump’s tariffs may have plateaued, as December net customs receipts totaled $27.9 billion, down from the low $30 billion range in recent months but far above ?the $6.8 billion recorded in December 2024.

Net customs receipts for the first three months of fiscal ?2026, which started October 1, totaled $90 billion compared to $20.8 billion in the prior-year period.
The Trump administration implemented some tariff-cutting trade deals in November, including 10 percentage-point reductions in duties on imports from China and South Korea. The Supreme Court also ?could soon rule on legal challenges to Trump’s tariffs under an emergency sanctions law. A ruling against those duties would further cut customs receipts.

The Treasury said that after making adjustments to December budget results in both 2024 and 2025, the December deficit would have been $112 ?billion, a decrease of $14 billion or 11% from the December 2024 budget gap.

Not to belabor the point but large deficits are a foreseeable consequence of cutting taxes without politically or structurally credible mechanisms for trimming expenditures. With entitlements on demographic autopilot, interest expense rising, and defense spending politically insulated, the notion that tax cuts would be offset by spending restraint was always fanciful. The pre-existing debt overhang already impedes economic growth by crowding out private investment and increasing the sensitivity of the budget to interest-rate shocks so you can’t expect to grow your way out of the issue. That is an empirical finding by economists that remains robust in mainstream literature.

These fiscal realities are not an abstraction. They are directly related to what Democrats frame as their “killer issue” in the 2026 and, possibly, 2028 elections: affordability.

When deficits are effectively monetized, i.e. financed through central bank balance sheet expansion rather than real resource growth, the result is inflation. When goods and services inflation outpaces wage inflation you have an affordability problem. It is always regressive: it hits renters, younger households, and fixed-income retirees first and hardest.

This is not a pathology unique to one party; it is the product of decades of bipartisan normalization of deficit finance in the service of short-term political objectives.

It has been a week since Maduro was apprehended and taken to the U. S. for trial. President Trump said the U. S. would “run Venezuela”.

Who’s running Venezuela? It appears to be the same regime that did a week ago with all of the same problems and failures.

I have seen some reports that some political prisoners have been released.

Everything seems to be proceeding much as I predicted.

If David Ignatius, the Washington Post columnist, still represents the voice of conventional Washington wisdom, they severely misunderstand what is happening in Iran. In that light his most recent column, considering Iran, reflects that misunderstanding sharply. He opens with a reflection on the Iranian regime and continues by characterizing it:

The Iranian regime is on a one-way street to disaster. A senior European diplomat in Tehran shared that assessment with me several years ago, and it remains true. Iran has powerful security tools, but they’re getting rusty. The regime couldn’t protect its proxies Hamas in Gaza, Hezbollah in Lebanon, and President Bashar al-Assad in Syria. And most important, Iran couldn’t shield itself from Israel’s systematic assault in June. The regime is on a losing streak.

“The Islamic Republic is today a zombie regime,” argues Karim Sadjadpour of the Carnegie Endowment for International Peace. “Its legitimacy, ideology, economy and leader are dead or dying. What keeps it alive is lethal force. It kills to live and lives to kill. Brutality can delay the regime’s funeral, but it can’t restore the pulse.”

concluding:

The wild card this year is whether the regime’s hard-liners have lost their edge. Like the Soviet Union during its last years, the security agencies may have lost their ideological commitment and discipline. They’ve watched helplessly as their proxy forces were crushed in Gaza, Lebanon and Syria. And they’ve suffered the same scourge of inflation and economic stagnation as the rest of the nation. They’re not broken, but they appear more fragile than in the past.

Let’s think again about what’s going through the heads of those 12 members of the Supreme National Security Council. They know Iran is stagnant. Their budget for the next fiscal year will likely raise taxes and cut subsidies to fund defense spending. They’re waiting for the political transition that will come when Khamenei dies, but for the moment they are lurching forward as a wounded dispirited nation.

As the council members look south across the Gulf to Saudi Arabia and the United Arab Emirates, they see their neighbors racing to embrace artificial intelligence and a dynamic economic future. Meanwhile, they cling to a repressive, retrograde regime that can barely feed its people. Revolutions are impossible to predict. But when you look at Iran, it’s obvious that eventually something is going to crack wide open.

The highlight is mine. I’m afraid he has several misconceptions. The Iranian regime is theocratic. It bases its authority on religious beliefs not popular support. In their own light they are not decaying institutions; they are custodians of divine mandate. As long as Iran is governed by True Believers, they will hold onto power by whatever means necessary. They must. Secular Westerners find that very difficult to understand.

Also, the analogy he considers in the body of the column, to the last days of the Soviet Union in the passage highlighted above, is correct but he does not take it far enough. A Marxist regime can drift into cynicism. A theocracy cannot. When belief dies, so does authority. Until then, repression is not a policy choice it is a theological obligation. The Soviet Union persisted until it was no longer governed by anyone who remembered the revolution. When that happened it collapsed rapidly.

The Soviet gerontocracy governed out of habit. The Iranian leadership governs out of covenant.

Sadly, Iran is still governed by actual religious revolutionaries. The question is not whether they will suppress unrest but whether they can.

I was composing a longer post on the situation in Iran but I just wanted to quickly make a few points.

There is a growing tendency to claim that “Iranians are not Muslims.” That strikes me as analytically sloppy. Opposition to the clerical regime is not the same thing as abandonment of Islam. One can plausibly argue that most Iranians are not Khomeinists; it is far less plausible that they are not Muslims. A few online polls don’t tell us much. Those notoriously oversample people who are young, educated, urban, and tech-savvy. That does not describe anything like the majority of Iranians.

Second, I don’t think we really know what’s going on in Iran right now. If the violent protests are large enough, they’re probably visible from orbit. Western human intelligence in Iran has been degraded over the period of years and much of what we know is inference layered on social media noise.

Third and this didn’t occur to me until recently, one factor that may matter more than is being discussed is the effect of Israel’s war with Hamas and Hezbollah on the Basij militia. The regime’s internal security apparatus does not exist in a vacuum; it draws legitimacy, manpower, and ideological energy from the same ecosystem that has been mobilized externally. Whether that has strengthened the Basij through mobilization or weakened it through distraction and attrition is an open question but it is unlikely to be neutral.

Megan McArdle lays out a very good case in her column in the Washington Post on the killing of a woman in Minneapolis by an ICE agent and the subsequent feces-flinging contest. She opens with a warning about police states and follows that with this anecdote:

In 2012, a group of law professors published the results of an experiment they had run on 202 adults who were shown a video of protesters. Participants were given the text of a law regarding protests at sensitive facilities and asked to determine whether the police had been justified in shutting down the protest.

Half were told that the video showed pro-life demonstrators at an abortion clinic. The other half were told the protest occurred outside a college career-placement office where military recruiters were conducting interviews, and that the protesters were rallying against the military’s “don’t ask, don’t tell” policy for gay and lesbian service members. The results were depressing, if not entirely surprising.

People disposed to support abortion rights and oppose “don’t ask, don’t tell” thought the police were justified in clearing protesters away from the abortion clinic but not the recruitment office. Those whose views went the other way reached the opposite conclusion from the same facts.

She then contradicts her own point by trying to analyze the various videos of the events.

The only pertinent question under the law is what was the state of mind of the ICE officer who fired the shots? The killing was unmerciful and unjust but was it illegal? That is entirely dependent on the state of mind of the ICE officer and the only evidence we have of that is his own testimony. It cannot be inferred from videos.

That is why I have taken the position that I have: that the officer should be prosecuted to the full extent of the law, that he is likely to be convicted because the jury is composed of people who will behave just as the “202 adults” did, viewing the events through their own peculiar prisms rather than as “reasonable persons”, and that the conviction is likely to be overturned on appeal.

We have built a legal system that depends on rational actors, and a political culture that ensures they do not exist.