
Something consequential is happening to democratic leadership. Not a scandal. Not a collapse. A developmental inversion. It is Smooth cognition under authority conditions.
Systems designed to surface judgment now reliably eliminate it. The traits once required to hold authority under uncertainty—absorbing dissent, tolerating ambiguity, standing publicly without resolution—are no longer selected for. They are filtered out early, quietly, and often with applause.
What’s left isn’t stupidity—it’s just a different setup. These are leaders who do great when everything’s stable. They’re smooth, quick to adapt, and tuned in—as long as someone’s showing them the way. But they’ve never had to fly solo, so they just don’t know how.
This is not a political argument. It is not ideological. It is developmental. A structural gap now exists between what governance requires during rupture and what the leadership selection system can produce.
The terms come from the companion page Smooths and Jags, and they describe a direction of movement, not two kinds of people. Most of us carry some of each. This page asks a narrower question: what happens when the systems that choose our leaders reward one direction, year after year, over the other?
This page calls the gap between what governance requires and what the selection system can produce the leadership void.

The remaining Jagged figures in national leadership share a pattern that resists spin.
They are old, or they were shaped by hardships most children today will never face.
Bernie Sanders, born in 1941, and Elizabeth Warren, born in 1949, grew up before everything was smoothed. They came of age when feedback was slow, questions stayed open for months, and no one came to rescue them. Both spent years in public holding positions that weren’t winning: Sanders argued for single-payer health care for decades, and Warren pressed for a consumer financial protection agency long before one existed. Holding an unresolved position in public wasn’t a choice; it was the job. That weight stayed with them.
John Fetterman may be a Jag of a different kind. A stroke in May 2022, days before he won Pennsylvania’s Senate primary, left him with auditory processing difficulties he has discussed openly; in his October 2022 debate, he followed the questions through closed captioning. Since then, his speech has been slower and less fluent, with pauses left standing. Whatever one thinks of his politics, that is hard to smooth and hard to mirror. What reaches the public sounds more like someone working out an answer in real time.
J.D. Vance, born in 1984, offers a younger version. In his 2016 memoir, Hillbilly Elegy, he described an unstable childhood in Ohio, much of it spent in his grandparents’ care, before he enlisted in the Marine Corps. That history doesn’t guarantee judgment under pressure, but it is the kind of ground where the capacity this page calls Manual Mode can form.
These are not endorsements. They are observations about formation, drawn from public record.
Fast forward twenty years. Nobody in charge will have grown up outside these friction-free systems. Kids today are raised where any confusion gets solved right away, feelings get smoothed over before they even bubble up, and struggling to think is treated like a problem to be fixed. When something bad happens, people handle it fast so it doesn’t leave a mark—or build any grit.
The pipeline that once accidentally produced Jagged leadership closes without notice. Not because society rejected such leaders, but because it eliminated the conditions required to produce them.
Smoothness Under Pressure
This is where the framework becomes predictive instead of just descriptive.
Some modern candidates feel “Jagged” to opponents because they are ideologically disruptive. But in how they perform under pressure, they can be Smooth: shaped by years of instant feedback to answer immediately, stay on message, and never be caught without a frame.
Marco Rubio is Smooth in the traditional sense: disciplined, message-tight, and rarely off script. He also shows how modern Smoothness develops, because his public record marks a clear before and after. In the Republican presidential debate in Manchester, New Hampshire, on February 6, 2016, he repeated a nearly identical line about President Obama several times, even after Chris Christie called it out on stage as a “memorized 25-second speech.” Coverage the next day described him as stuck in a loop. It was a rare public moment when the polish showed its seams. A decade later, as a senator and then as Secretary of State, Rubio’s public performance has grown noticeably more disciplined, and a moment like Manchester is hard to imagine from him now. Whether AI plays any part in that, the public record doesn’t say. What has changed is what any campaign or office can now do: test phrasing in advance, rehearse against simulated attacks, and catch a loop like that one long before it reaches a stage. The slip that exposed him in 2016 is exactly the kind of friction the new tools are built to remove.
The record also contains a more recent counterexample. On June 3, 2026, appearing before the Senate Foreign Relations Committee, Secretary of State Rubio faced pointed questioning from Senator Jeanne Shaheen over the administration’s temporary easing of Russian oil sanctions. She pressed him repeatedly for a commitment not to extend the waiver. Rubio responded by distinguishing between global energy supplies, Treasury’s authority, and the conditions under which the waiver might expire, and he declined to make the commitment she wanted. The answers can be challenged on their merits, but the exchange didn’t resemble Manchester: faced with interruptions and follow-up questions, Rubio kept responding to the substance rather than returning to a memorized phrase.
The contrast matters. The 2016 failure was real, but it never proved Rubio couldn't think on his feet. And that is the harder point. Polish no longer tells us which kind of leader we are watching. A disciplined performance can rest on real judgment or stand in for it, and from the outside, the two now look the same.

Alexandria Ocasio-Cortez offers a different kind of example. She is Smooth in the modern sense: rhetorically fluent, native to social media, emotionally expressive without routinely losing control of the message, and skilled at compressing complicated political arguments into memorable moral frames. Her statements are often built for the short clip as much as the long exchange. Opponents may see her as Jagged because she challenges established positions. But disrupting someone else’s political narrative is not the same as jaggedness in one’s own communication.
On 60 Minutes on January 6, 2019, responding to criticism of factual errors, she said: “I think that there’s a lot of people more concerned about being precisely, factually, and semantically correct than about being morally right.” She went on to argue that her mistakes should not be equated with deliberate falsehoods, and she acknowledged that factual accuracy mattered. Read in full, the remark is a defense, not a confession. But it names a tension this page worries about: the pull of a clear moral frame even when the underlying facts are disputed. It was also a line built to travel. Supporters and critics repeated it far beyond the interview, often without its context.
The record also contains a counterexample from the same year. At an MSNBC town hall on climate change on March 29, 2019, an audience member shouted “Moron!” at Bob Inglis, a former Republican congressman on the panel who had argued for a more gradual approach. The insult came from an audience largely sympathetic to Ocasio-Cortez, and the easy move was to let it pass. Instead, she interrupted: “Hey, hey, hey, hey, that’s unacceptable.” Then she added, “And that’s the difference between me and Trump.” Defending an opponent against her own sympathetic audience, live and with no apparent script, meant breaking the room’s momentum rather than riding it.
There is a complication worth keeping. Even that spontaneous moment produced a portable political line: she defended an opponent and, in the same breath, drew a memorable contrast with Trump. An authentic response can also be rhetorically effective. Both can be true at once.
As with Rubio, the record shows both: the clip-ready moral frame and the moment she challenged her own audience. Neither suggests her communication is artificial or AI-assisted. The larger question is what happens as technology learns to manufacture the outward signs of fluent, controlled, emotionally resonant speech. The qualities that once suggested a person was thinking clearly in the moment can now be rehearsed, optimized, and reproduced. What gets harder to see is whether the polish reflects independent judgment or only its successful presentation.
These moments are not verdicts on either person. They are public records of how polish behaves under pressure, and the reading offered here is interpretation, not diagnosis.
Smooth leaders are not new. What is new is how hard it has become to see the seam, the place where performance ends and judgment begins, from the outside.
The point is not whether Smooth leaders are “bad.”
Smooth leaders can be intelligent, competent, and sincere.
The point is that AI is likely to amplify Smoothness as a political survival trait.
This suggests the political ecosystem will increasingly favor people who:
In an AI-mediated environment, that kind of leader can survive indefinitely—because the friction signals that used to expose weakness can be smoothed away before the public ever sees them.
The real question is no longer Smooth vs Jagged.
The real question is:
Do they still have Manual Mode?
Can they think under load?
Can they hold ambiguity?
Can they resist premature closure?
Can they govern without a script?
Because Smoothness is safe when it is backed by real cognition.
Smoothness becomes dangerous when it becomes a performance stabilized by machines


Democratic systems did not always select this way. The inversion occurred in stages.
First came feedback acceleration. Polling tightened. Messaging shortened. Response cycles compressed. Leaders learned that hesitation signaled weakness and delay signaled incompetence. Judgment became something to display instantly, not something to carry.
Then metrics took over. Approval curves. Engagement graphs. Donor signals. Performance became legible only through numbers. Anything that could not be measured became suspect. Anything that could not be optimized became invisible.
Algorithmic amplification followed. Messages that connected spread faster. Resonance became confused with truth. Fluency with virtue. Leaders who spoke cleanly inside existing frames were rewarded. Those who introduced friction—new frames, unresolved questions, uncomfortable pauses—were punished by silence.
Finally, institutions internalized the pressure. Parties, media, staffers, consultants began pre-filtering candidates. Not maliciously. Practically. Why advance someone who “doesn’t land”? Why risk ambiguity when clarity performs better?
Independent judgment—once the point of leadership—became the trait most reliably eliminated by the process meant to find it.
This is the inversion.
A final stage is now arriving — and it changes the selection system again.
AI does not just reward Smoothness.
It turbocharges it.
In the past, even the best Smooth leaders had limits. They got tired. They misspoke. They hesitated. They lost control in a debate. They showed friction. And that friction was a signal — not of weakness, but of reality.
AI removes that signal.
A candidate can now pre-write answers, pre-test phrases, simulate debate attacks, and generate “presidential language” on demand. They can deliver perfect empathy. Perfect framing. Perfect closure. And they can do it every day without the fatigue that used to expose whether the person underneath could actually govern.
This is no longer hypothetical. Campaigns have already begun using AI to draft political messages, test their effectiveness, and extend a candidate’s voice beyond the candidate. In Virginia’s 2023 legislative elections, Tech for Campaigns, an organization supporting Democratic candidates, tested fundraising emails written with Google’s Bard against emails written by people. According to Axios on January 30, 2024, the AI-assisted emails raised 3.5 times more money per hour of work. Human editors still reviewed the material, but the economics had changed: the machine produced more language faster, and the campaign measured which versions worked. The Associated Press later documented AI’s use across Democratic campaign operations, including email drafting, voter targeting, and data analysis. Smoothness was becoming something that could be produced, compared, and optimized.
The technology has also moved beyond drafting. In October 2025, Virginia Republican lieutenant governor candidate John Reid staged a debate against an AI-generated version of his Democratic opponent, Ghazala Hashmi, after she declined a conventional debate. The simulation was not Hashmi, and its answers could not establish what she would actually have said. But the spectacle showed a new capability: a political exchange could be manufactured without one of its participants. By October 2026, the Center for Democracy and Technology was documenting AI-powered voter outreach systems able to carry on individualized text conversations. Political communication was no longer limited to what candidates and their staff could personally write, rehearse, or deliver.
What is new is not the coaching. Speechwriters, debate rehearsals, focus groups, and message discipline are as old as modern campaigning. What is new is the scale, the speed, and the stamina. A human consultant can rehearse a difficult question dozens of times. A machine can generate hundreds of variations, simulate objections, and repeat the exercise without fatigue. That doesn’t mean the candidate will never stumble, or that AI-generated language will always be right. It means the machinery around the candidate can sustain a level of polish that once required a far larger human operation. The candidate can still slip. The messaging system no longer has to.
This is not “AI replacing leaders.”
This is AI stabilizing the performance layer so completely that the public may never see the difference between real judgment and performed judgment.
That is the new danger.
Smoothness used to be a style.
Now it can be an externally maintained condition.
And the more politics becomes a fluency contest, the more the system selects for leaders who can speak in finished paragraphs — even when their internal cognition cannot operate without scaffolding.
This is Synthetic Competence at the level of authority.

Smooth leadership isn't clueless. It actually works well when things are steady—when you know what's going on, get quick feedback, and always have support. The cracks only show when things fall apart.
At some point, the loop sealed itself.
A Smooth electorate selects Smooth leaders. Those leaders produce Smooth policy—policy optimized for clarity, predictability, and immediate legibility. Those policies generate Smooth systems: education that resolves early, media that explains instantly, technology that completes thoughts before they fully form.
Citizens raised inside those systems experience friction as harm. They reward leaders who remove it. The system feeds itself.
There is no internal exit.
This becomes obvious in AI regulation. Legislators regulate what Smooth cognition experiences as harm: misinformation, deepfakes, privacy violations. These are loud failures. They disrupt coherence.
What Smooth cognition experiences as helpful—frictionless completion, emotional availability, instant guidance, preemptive rescue—does not register as harm. It registers as care. As progress. As safety.
The developmental damage documented by this archive does not appear on the radar of scaffold-dependent cognition. Smooth leaders will protect smoothness while sincerely believing they are protecting the public.
Then comes the dependency trap. Leaders now require AI tools to remain competitive. Intelligence synthesis. Diplomatic modeling. Military coordination. Economic forecasting. A leader without AI assistance is outpaced immediately.
But each successful use installs reliance. Judgment migrates outward not from laziness, but from efficiency. Over time, leaders lose the capacity to operate without the tools that made them effective.
The competitive advantage becomes the vulnerability.
There is a second way the loop tightens, and it runs through money. The companion page Smooths and Jags predicts private schools that deliberately preserve friction, for families who can afford them. If that happens, the capacity this page says leadership needs will not disappear evenly. It will concentrate in the families who could pay to keep it. A democracy could end up with judgment sorted by class: a few leaders formed in friction, drawn from a narrow slice of the country, governing a public that was never given the chance to build the same capacity.
Now imagine two Smooth-led nations in conflict. Both AI-dependent. Both scaffold-reliant. A system failure, data poisoning, or adversarial manipulation disrupts the AI layer. Signals contradict. Models diverge.
Neither side has leadership capable of Manual Mode.
Escalation follows—not from aggression, but from the absence of authorship. Decisions are made to restore coherence, not to track reality. Both sides scan for signals that no longer exist. This is a scenario, not a forecast. It describes a failure mode worth planning against, not an outcome anyone can predict.
The leadership void does not look like chaos. It looks like procedure without movement.
The room fills with capable people. They speak fluently. They update each other. They repeat known facts. Statements multiply. Nothing advances.
This is not an argument. It is a stall.
Everyone waits for clarity that cannot arrive because the machinery that produces clarity is itself incoherent. Language accelerates to cover the free fall. Committees form reflexively. Narratives are proposed to quiet the discomfort.
Look at what reaches the head of the table. In the hours before the meeting, staff have used AI to summarize the intelligence, draft the options paper, and prepare talking points for each choice. Every document is clear. Each option includes a rationale, a risk estimate, and a recommended next step. What none of them contain is the analyst’s doubt, the field report that didn’t fit, or the junior officer’s sense that the situation isn't what it looks like. Those were the rough edges, and the drafting removed them. The leader at the head of the table may have real judgment. It may not help much, because the friction that judgment needs was filtered out before the briefing book was printed.
No one wants to be the first to say, “I don’t know what this is.” That sentence no longer functions as leadership. It functions as exposure.
So the room waits. Actively. Scanning for any signal— even a wrong one—because wrong with coherence feels safer than right without it.
This is the void: a room full of adults who cannot move until someone tells them what reality is—and no longer know how to recognize it themselves.


No transformation is on offer here. Manual Mode is built through years of friction, most of it in childhood, and nothing in this archive shows it can be installed in an adult on demand. A Smooth leader does not become Jagged through insight, a briefing, or a leadership course. Some adults rebuild part of the capacity, slowly, through deliberate effort over years. A crisis will not wait for that.
This is harm reduction.
Some leaders pre-commit to slower authority. A rule written in advance: no irreversible decision in the first hour of a novel crisis. Not because waiting produces wisdom, but because premature coherence produces damage.
Some designate friction-speakers ahead of time. Not advisors. Not strategists. People whose sole function is to say, “This doesn’t resolve,” without punishment. This delegation must occur before crisis. In the moment, it will not be chosen.
Some practice unknowing publicly. Naming ambiguity in low-stakes settings. Saying “we don’t know yet” while still occupying authority. This reduces limbic shock when ambiguity arrives uninvited.
Some rehearse silence. Ending sentences early. Letting pauses stand. Unpracticed, this feels like loss of control. Practiced, it becomes survivable. Silence is not wisdom. It is containment.
For this moment, the four practices above belong on an emergency card, written in calm and kept within reach of anyone who may have to decide under pressure: no irreversible decision in the first hour, the friction-speaker named in advance, “we don’t know yet” said aloud, and silence allowed to stand. The card is not guidance. It is gravity. It reminds the leader that eloquence is now a liability. That quiet may prevent irreversible harm. That deferring to someone who can operate without signal is not weakness—it is damage control.
That is the ceiling

The leadership void is not a crisis to be solved. It is a condition already in effect.
The system will not self-correct. The selection pressure is stable. The loop is closed. Leaders will continue to be chosen for fluency, responsiveness, and alignment. The traits required to govern under rupture will continue to age out without replacement.
What remains is limited and unsentimental: individual containment, private transmission, institutional pockets that tolerate friction, and documentation for reconstruction.
The archive does not prescribe.
It records.
This page is an argument, not a study. The research beneath it is collected on the companion page Smooths and Jags, including Sparrow, Liu and Wegner (2011) on memory and search, Bastani and colleagues (2025) on AI-assisted practice and unassisted exam performance, and Gerlich (2025) on AI use and critical thinking. Those studies measure individuals on tasks. None measure political leaders, so the step from their findings to leadership is my inference. The public record in Sections II and III, from biography to the debate, hearing, interview, and town hall cited there, is linked to its dates and sources, and my reading of it is interpretation. The conflict between two AI-dependent nations in Section V is a scenario, not a forecast.
ARCHIVAL SEAL
© 2026 Jim Germer · The Human Choice Company LLC. All Rights Reserved.
We use cookies to improve your experience and understand how our content is used. Nothing personal -- just helping the site run better.