Something peculiar is happening here.

AI labs are writing constitutions for machines while governments are fighting over whether they themselves still get to write the rules. Anthropic has Claude's Constitution. Amanda Askell gives the thing virtues, character, and "hard constraints." Jill Lepore, writing in The New Yorker, asks the obvious question: what does it mean when a private company starts sounding more constitution-minded than the constitutional order around it?

At the same time, the federal government wants one national AI framework and has argued that state laws are an obstacle to innovation, while California and other states keep trying to impose safety, privacy, and reporting guardrails anyway. So this is not a story about "whether governance exists." Governance exists. It is simply fragmented, contradictory, and fighting with itself in public.

And then there is the more antagonizing part: the discourse of alignment somehow feels like the main question is whether the model is "good", whatever the heck that means. But lurking just below that is the real thorny (intractable?) question — who (what?) has authority? Who sets the red lines. Who decides what surveillance is allowed. What decides is a right answer or plan or guidance when a question is asked of a machine intelligence? Who gets to build these intelligences (whether a model or the things that wrap around them or the devices/services that rely on them) and are they in any way regulated the way we might regulate food, building, medicine, or sandwich shop for the good of the public? Who decides whether autonomous weapons are acceptable. Who decides whether a private model vendor, a procurement office, a governor, a legislature, or the Pentagon gets the last word?

Recent reporting has made this whole show basically impossible to ignore. There is the White House push for federal pre-emption. There are the state fights over guardrails, access, and privacy. There is the Anthropic conflict around surveillance and autonomous-weapons boundaries (as if a private company should be the one drawing the lines around what counts as acceptable surveillance or autonomous-weapons behavior — remember Google promised to do no evil once..and changed its mind once the PR value of that promise shifted). There is Ezra Klein talking with Ben Buchanan about AGI preparedness as a state-capacity problem rather than a mere product launch problem. And beneath all of this is a much larger issue: not just how powerful these systems become, but what kind of institutional order is being built around them.

Access, for one thing, is becoming constitutional.

Who gets the good model? Who gets the throttled one? Who gets denied? Premium tiers, restricted compute, export controls, enterprise procurement filters, national-security exceptions, and platform-controlled inference all point to the same thing: whatever "democratic access" to AI is supposed to mean, it is not likely to arrive as some evenly distributed civic blessing. It will be rationed, credentialed, subsidized, denied, and politically narrated. That is not a bug in the story. That is the story.

And if the liberal-democratic constitution is weak, other constitutions rush in. Managerial constitutions. Security-state constitutions. Corporate constitutions. Technocratic constitutions. Perhaps even the anti-constitution: not rights, not representation, not public contestation, but optimization, expertise, performance thresholds, and restricted participation for the supposedly qualified few. So then the question becomes: what does a technocratic order actually look like in everyday life? What badges, notices, clearances, memos, dashboards, handbooks, and eligibility forms tell you that governance has shifted from public legitimacy to expert administration?

This is where General Seminar does what General Seminar does best.

We are not going to sit around and perform AI ethics. We are going to turn the current argument into legible artifacts from inside these possible worlds. We will make the governance struggle visible as a procurement red-line memo, an AI vendor safety disclosure sheet, a state-federal conflict map, an internal policy brief for leadership, a legislative hearing question pack, an incident report from a near-future governance failure, a Right to Access AI bill one-pager, an inference rationing notice, an AI access-tier chart, an anti-constitution manifesto, a technocratic operating charter, a model-governance credential badge, or an expert-council memo authorizing restricted decision rights.

That is the move here: not AGI prophecy, but institutional preparation. Not ethics theater, but governance rehearsal. Not hand-wringing over whether the future will be strange, but making the strangeness tangible enough that policy people, strategists, and institutional actors can actually do something with it.

If your job is to brief leadership, write policy, shape organizational strategy, evaluate risk, or explain to others what kind of AI order is taking shape around us, this General Seminar is for you.

Get your ticket now.

Produced right here in Venice Beach California by Near Future Laboratory

Learn how to get reimbursed for this professional learning & development package!