Dispatch: The Manhattan Primary is Referendum on AI Governance
So, here’s a wild rabbit hole I’ve been going down lately: the NY-12 House primary.
In any normal time, the story would be the institutional inheritance fight. Micah Lasher — Nadler’s longtime aide, Bloomberg’s former state director, now backed by both — is the heir apparent in the most literal sense. Jack Schlossberg, Caroline Kennedy’s son, JFK’s grandson, and social-media-forward millennial, carries a different kind of inheritance. George Conway, meanwhile, is the witty Never-Trump Republican who became a Democrat to run for this seat. Who gets the great old liberal institutional seat in Manhattan? Staffer, scion, or convert?
But that’s the sideshow.
The race is becoming an early referendum on AI governance.
Alex Bores — assembly member, former Palantir staffer (he resigned in protest), master’s in computer science — is the lead Assembly sponsor of the RAISE Act, which Governor Hochul signed in December over fierce industry opposition and despite a Trump executive order attempting to block state AI regulation in the absence of a federal framework. The law requires the largest frontier AI developers to publish safety protocols and disclose serious incidents. It is, at present, the strongest AI transparency statute in the country.
That has made Bores a target. A bipartisan Super PAC called Leading the Future, launched last August with more than $100 million, announced in November that it would spend millions specifically to sink his congressional bid. There is a real argument for federal preemption — fifty different state regimes for frontier models will be a mess to manage, costly to national competitiveness, and a coherent national framework would be better than what we have. But that argument lands differently when the same actors opposing state action are not putting forward other ways to solve the governance challenge that would be legitimate with the public.
All of this is unfolding while, in an Oakland federal courtroom this week, Elon Musk is on the stand testifying that Sam Altman and OpenAI “stole a charity.” Musk argues that OpenAI betrayed its founding nonprofit mission to develop AI for the benefit of humanity. OpenAI argues its structure was the only way to marshal the resources frontier AI requires, and that Musk’s suit is shaped by his commercial rivalry through xAI.
Whatever the merits, the trial puts the deeper question plainly: who gets to govern technologies this powerful? What incentives should drive their development?
The Musk-Altman trial will be decided by a federal judge in Oakland. The RAISE Act was decided in Albany. The federal preemption attempt — made by executive order rather than legislation — bypassed the venue actually designed for this question. The counter-pressure is being applied by a $100 million Super PAC. None of these venues is broadly democratically legitimate, or likely to build anything like a public mandate.
And so, here we are. A local congressional primary in Manhattan has become a proxy fight over whether democratic institutions can meaningfully shape AI before AI companies shape the institutions that govern them. And if they can, at what level. District? State? National? Through some venue or instrument we haven’t yet built?
Jerry Nadler’s career belonged to an era of liberal constitutional politics — civil liberties, impeachment, courts, rights, institutions, the slow machinery of law. The race to replace him belongs to AI, platform power, outside money, and a question Nadler’s era never quite had to answer: whether democratic institutions can move fast enough to govern technologies that are already reorganizing the society those institutions are meant to serve.
What I keep coming back to is that the institutional order Nadler spent his career inside is not waiting for AI to finish reorganizing it. It is, in many ways, already gone. And the race in NY-12 will be an early indicator of what we build next — who does the building, according to what principles, and rooted in what legitimacy?
Up Next:
Observant readers will notice I’ve been quiet lately. Why? Some combination of parenting a toddler, work on other projects, and tortured wrestling with a piece on political alienation. If you have thoughts on that subject, please reach out. I’d love to have more folks to chew over it with.
What I’m reading:
I have finally finished the tome is that is the David McCullough Adam’s biography. It’s one of those books where finishing it feels like leaving your friends behind. What better company than Adams, Jefferson and Abigail? And McCullough is so generous but clear-eyed about humanity. Next up for me is McCullough’s follow up: 1776.



One key issue seems to be agility . Our governmental and legislative systems are like supertankers navigating the ocean , slow to turn and difficult to change course. AI behaves like a smart intelligent race boat , continually evolving and changing course navigating the rocks of regulation with ease. Until we match the speed of evolution with speed of regulation or something completely different to regulation
. AI or whoever controls it will win the race. The prize being an uncertain and random future outcome difficult to predict.