DON'T MISS: Our Publisher Pascal-Emmanuel Gobry has a new op-ed in the Washington Post on the administration's new H-1B visa fee policy.

What The Network State Movement Is And Why You Should Support It

Dear reader, your correspondent spent the past few days in Singapore, and in Forest City, Malaysia, a planned city outside Singapore, attending (and speaking at) the Network State Conference.

What is a "network state"? The name is the brainchild of famous tech investor Balaji Srinivasan. A network state is a project to create a new state, by first creating a network of like-minded people online, eventually establishing some sort of offline presence in the real world and, eventually, turning that place into a new independent microstate.

Does it sound crazy to you? Well, it might.

But we would say a few things.

The first is that everyone we met at the Network State conference was enormously impressive. Very serious people, from successful crypto entrepreneurs to large real estate developers, are gaining an interest in the Network State movement. These are, mostly, not cranks or crazy people, but serious people with serious resumes and serious capital at hand.

The second is that it's easy to see why there is demand for something like this: most states, particularly in the West, seem incompetent, exhausted, close to financial bankruptcy in spite of confiscatory taxes, and to lack the ability to ensure basic government functions.

More generally, governance today is just dreadfully boring. There is a weird sense in which, because of the internet, all Western countries are copying each other and implementing the same policies at the same time, whether it's euthanasia or trying to replace the energy mix with windmills. There seem to be only two competing government models: "liberal democracy," with all its myriad flaws, and the Chinese combination of one-party totalitarian state and state capitalism, with all its myriad flaws too.

Well, how are the network staters doing? Well, some of them we spoke to have advanced quite far: they have built cities, or at least residential and business properties, in some piece of the world, and are attracting like-minded people to live there. Some of them have agreements with their host governments for special benefits, such as eligibility to be a "Special Economic Zone." Some are still just at the planning or theory stage—which is also ok.

We believe the network state movement has some real wind in its sails. As time passes more and more people will look to escape the dark horizon of Western states, and not just to cyberspace, but physically.

The fundamental issue is this: competition and innovation in governance is good. There should be more city states. There should be more states with all sorts of political regimes. Let a thousand flowers bloom!

You will hear more of the network state movement, and you should support them.

Policy News You Need To Know

#AI — In a recent piece, AEI Senior Fellow Will Rinehart takes aim at the increasingly popular practice of AI researchers sharing their "p(doom) estimates," that is to say, the probability that artificial general intelligence will cause catastrophic harm to humanity. While figures like Geoffrey Hinton place their estimate between 10-50% and Eliezer Yudkowsky above 95%, and Marc Andreessen at 0, Rinehart argues that these unconditional probability statements are fundamentally unhelpful for actual policymaking. Drawing on his background in Bayesian econometrics, he contends that meaningful risk assessment requires specifying conditional probabilities—how various interventions, governance decisions, and safety measures would alter the likelihood of catastrophic outcomes. Rinehart is particularly critical of Yudkowsky's framework, which posits that doom is both inevitable and incomprehensible (involving scenarios like superintelligent "dust mites"), arguing his approach puts the argument "beyond the reach of rational debate" and sidesteps the hard work of modeling specific failure modes and developing targeted interventions. Instead of trading p(doom) numbers like "restaurant recommendations," Rinehart urges the AI safety community to focus on actionable questions: under what conditions does risk increase or decrease, and which interventions offer the highest expected value for reducing catastrophic outcomes—the difference, as he puts it, between paralyzing fatalism and productive risk management. We agree and found his explanations very valuable.

#Shutdown — Michael Strain from AEI makes a compelling case in today's FT that Democrats have badly miscalculated their shutdown strategy over ACA subsidies. His core argument is straightforward: House Republicans have passed a clean continuing resolution funding the government through November 21 at current Biden-era levels—a position he characterizes as "shockingly reasonable"—yet eight Senate Democrats refuse to break ranks and pass it, effectively owning the shutdown despite early polling showing Trump taking more blame. Strain contends that Democrats fundamentally lack leverage here because previous attempts to use shutdowns to force unrelated policy wins have consistently failed, and ironically, the shutdown is actually strengthening Trump's hand by allowing him to implement spending cuts, hold funds for New York infrastructure and climate projects, and potentially advance workforce reductions. He predicts the standoff will be short-lived—probably ending by October 15 when military paychecks are due—and argues that Democrats will likely secure some form of ACA subsidy extension anyway once they reopen the government, making their current strategy both ineffective and damaging to US economic data collection at a critical moment when the Fed is misreading inflation signals. Worth noting that the FT comments section lit up with accusations of partisanship, but Strain's analysis of Democratic leverage (or lack thereof) seems pretty sound.

#FertilityCrisis — Finally some good news. The latest Pew survey shows a significant uptick in Americans recognizing the negative impact of fewer people choosing to have children. It's now a majority, 53%, up 6 points from last year. Republicans lead at 63% and men show particularly strong gains. What's especially notable is the 7-point jump among Democrats to 44%, suggesting this is becoming less of a partisan wedge issue. While a majority still resists federal intervention (56% say government should have no role), those who support government action overwhelmingly favor tax credits for parents at 82%, a promising area for future policy movement. The partisan gaps emerge on mandates like paid family leave and free childcare, where Democrats predictably favor more government involvement.

#K12 — New research from AEI confirms something which we already know, but it's better to have strong numbers: that chronic absenteeism deserves its place at the top of state education policy priorities. While the pandemic-era investment in educational technology has modestly reduced the per-absence harm to student achievement—with each 10 absences now reducing math scores by about 6% rather than 7% of a standard deviation—this marginal improvement has been swamped by a 65% increase in total absences since Covid. The study, using administrative data from hundreds of thousands of students across Maryland and North Carolina, finds that absences continue to harm achievement in an approximately linear fashion, meaning every missed day matters and there's no safe threshold below the federal chronic absence definition. Notably, the harm from absences was similar across socioeconomic backgrounds, though high-performing students were hit slightly harder, suggesting that technology alone cannot substitute for physical attendance.

#EconomicPolicy — Paul Winfree, president of the Economic Policy Innovation Center and former Trump budget official, is in NR with a standard anti-QE argument. While the argument itself is not new, he brings new research to bear on the issue. While Winfree's research finds that QE doesn't directly cause fiscal profligacy during stable periods, he argues that it significantly amplifies Congress's temptation to overspend during crises, pointing to the COVID-19 stimulus as a prime example that delivered "the highest, most persistent inflation in four decades." The core problem, as Treasury Secretary Scott Bessent has termed it with his "gain of function monetary policy" phrase, is that massive Fed bond purchases mute the market signals that would normally discipline government borrowing through higher interest rates.

#IndustrialPolicy #AmericanManufacturing — New research from Brookings estimates that the CHIPS and Science Act led to a 15-21k increase in US semiconductor jobs and 27-33k job increase in upstream input sectors and non-residential construction. More here.

#Woke #DEI #VibeShift — DOT on a hot streak of placing infrastructure projects under administrative review for DEI contracting requirements. Two projects for Chicago’s CTA totalling $1.2 billion are being investigated.

#Immigration #BigTech #VibeShift — Apple has pulled the "Iceblock" app which people used to report ICE raids. Of course, obstructing Federal agents is a crime. It's still nice to see they did it without protest.

#2A — Supreme Court to hear challenge to Hawaii's ban on concealed carry.

#FedNews — Trump's pick to lead the Bureau of Labor Statistics, E.J. Antoni, was withdrawn after failing to gain Senate support.