Immigration in the Age of AI Could Look Very Different

The host of Second Orders is unnamed. He is Singaporean, and he says so. Life in Singapore, he tells Garett Jones, can feel like the movie adaptation of Jones’s books, and as usual the books are more fun. Jones is a professor of economics at George Mason, formerly of the United States Senate, now chief economist at Blue Chip, a stablecoin rating agency. He calls Hive Mind, 10% Less Democracy, and The Culture Transplant his Singapore trilogy. The host wants the forward implications if Jones’s theory is right, not another round of critiques. What follows is Jones’s argument, reported as his.

Ritualized civil war

One great thing about democracy, Jones says, is that it is ritualized civil war. Everyone knows which day to show up. Instead of fighting, we count noses and decide how to move on.

Deep roots

Deep roots, in the papers he is using, come from the migration experiments after 1500. Economists watch nature’s experiments. Country of origin matters. Places that received many immigrants from Western Europe tended to do well; so did places that received many from East Asia, especially China. Places that received many from sub-Saharan Africa tended not to. He presents this as an empirical pattern, a broad generalization robust enough to name a theory. How the traits travel — culture, genes, a fusion — is unknown and beyond current tests. What is old is new again.

The host tries a tradeoff. If a country saves too little, take frugal migrants; if it saves too much, take spenders. Jones pushes back. On savings and patience, more is always better in the long run. There is no paradox of thrift at the horizon that should occupy ninety percent of our attention. No country ruined itself by saving too much. The fortunate fact, he says, is that the good traits correlate. Psychologists call it g. People who are smarter tend to be more patient, more cosmopolitan, more liberal in their attitudes, more pro-trade. Select on IQ or on thrift and the other goods tend to come along.

Country rankings persist. Over the last century or so they have not moved much. The big rank changes are people leaving communism, or other command economies, for something more laissez-faire. Poland after 1990 is the obvious picture.

Papa John’s theory

He has a name for the institutional claim: the Papa John’s theory. Better ingredients make better pizza. Better people make better policies; worse people make worse ones. Getting the raw ingredients right is first-order.

There is reverse causation, he grants. Good institutions lift the education and health of people who come from places where those systems are weak. He does not know that this second arrow is stronger. It is important.

The host offers an evolution analogy and a worry about selecting against intelligence. Jones’s reply is not to negatively select on IQ. Openness to many European origins, work he associates with his colleague Jonathan Schulz, predicted later American innovation. Hybrid vigor is a real result: more diversity can keep deleterious recessives down. Short-run conflict, in this telling, can be worth it.

Diasporas as instruments of home-country influence he generally finds weak. If you worry about the pull home, give citizenship faster. A green card or a passport weakens the tug. People like holding one.

Point system and replacement

The point system he would write is about two-thirds individual traits and one-third country of origin. That, he says, is the right dose of deep roots for a rich country. Poor countries should weight deep roots more. They should subsidize high deep-root immigration. If Madagascar decided the theory mattered, it would take almost anyone from China who would come, have children, and stay a few generations. The phrase he uses for that poor-country strategy is “great replacement” — “a truly great replacement,” a term of abuse in Europe that he wants to turn into a development program. China’s second continent, the book he cites, already has about a million Chinese in Africa.

The host asks about AI agents negotiating each migrant’s terms. Jones is skeptical. Three simple indices would be enough: national test scores, national savings, the inverse of corruption. A fancy model will not, in his view, beat that R-squared. You cannot fake a country’s résumé. People fake individual packets — padding, disability waivers, a five-hundred-page file. Even if an agent could watch six months of your mail, he says, it would still collapse onto a few predictive factors.

His book Hive Mind is the IQ book. IQ matters, in that story, mostly through externalities and spillovers, perhaps half of them institutional: smarter people build better institutions, which help everyone.

AI governance fork

Put AI next to national IQ and the fork is sharp. If poor countries can credibly hand a lot of governance to AI, he thinks you could see catch-up miracles within a decade. A couple of experimental middle-income or poor states will go full AI-max. If they cannot make that commitment credible, O-ring high technology — the work that needs high-quality labor all the way down — stays in high-test-score countries, and the gap widens.

A dictator who imports talent to build the system and then deports it is the obvious cheat. The “AI took my job” meme is already strong enough that the dictator would have to fight it. You need credible contracts.

The place premium is not only institutions. It is also cognitive culture. Jones is a fan of the late James Flynn, who documented rising scores and had a theory of what the twentieth century did to minds. Surrounding a worker in Lagos with chatbots that copy the intellectual weather of San Francisco does not, in Jones’s view, get you a San Francisco chip fab. The long chains are not in the chat.

After ASI, who delegates still matters. He watches “Whimo” robotaxis and the push to keep a third of grocery checkout counters human. Those carve-outs are costly O-rings. If the gap between machine governance and human governance is tenfold or a hundredfold, small human reservations still tax the system.

AI is already in democracy under the desk: people checking it in the meeting, folding it into the laptop. Formal justices-as-AIs he thinks unlikely — too on the nose. It may reduce conspiracy beliefs; there is already experimental work in that direction, which would be good for voter quality. It will not automatically lengthen time horizons unless some Denmark-style country outsources a couple of bureaucracies, gets a loud result, and forces imitation. More ballot boxes were already possible once the internet arrived. People wondered about a plebiscitary flood. Jones does not think it happened. Voters cognitively economize. They track sports more carefully than politics. One reading is that they do not want the other clowns voting every day either.

Cross-coalition bargaining will persist. Legislatures still have to make grand bargains across committees. AI may improve the deals. It does not abolish the need for them.

Stablecoins as exit

Stablecoins he reads as Hirschman’s exit, not voice. Voice is democracy, loosely. Exit is leaving the currency, or the country, without permission. In a middle-income state with a mediocre central bank, the dollar in software is a way out. He would not call that democratic. Live, on this show, he makes a new prediction he says he had not thought until now: the rise of stablecoins means lower inflation globally, and more small-country dollarization — four, five, maybe ten more Panamas.

Central bank independence is his working example of ten percent less democracy. Independent banks, in the Alesina–Summers picture he teaches, tend to have lower inflation without worse recessions. Poor countries have done better at this over the last decade. That, he says, is a reason to think some AI-led governance might stick. It is also the limit test the host uses: if stablecoins let citizens exit the national money, you have a form of voting with your wallet that is not a ballot.

The host thanks him. Jones thanks him for having him.