Collaboration with AI can take many forms
Could you tell us what role collaboration with AI played, in the production of this essay?
The central idea (avoid single-factor models, while acknowledging this is itself just a single factor among many) is entirely mine.
Several of the sub-threads, such as Structural Defenses (e.g. non-AI methods to prevent one-shottedness), the Activist-Ideologue distinction, and Metastasization of the Local Frame were suggested during discussion with Claude and GPT5 about the central idea, and I chose to incorporate them after further discussion. Many potential sub-threads did not make the cut.
Others of the sub-threads, such as Why This Persists, the Self-Own, and and the general importance of factors and difficulty of multi-factor thinking are my ideas that were deepened through discussion with AI.
Dimensionalization and Conceptual Rhyme are my ideas that I have written previously about how to get AI to perform on command; I developed these working closely with AI.
Some phrasing is AI-written, but most of the writing is my own.
This is the first essay in what I intend to be a brief series about using AI to help automate and scale perspective-taking.
TL;DR
Humans love single causes. We want one villain, one cure, one elegant model to explain everything.
But the world runs on interacting factors (economic, biological, cultural, psychological, physical, …). Collapsing that complexity produces zealotry, not understanding.
People get one-shotted all the time: by ideology, by drugs, by heartbreak, by a persuasive AI chat, by a tool that seems too powerful not to use. One overwhelming perception feels like final truth, and the rest of the worldview folds around it.
The antidote isn’t “it’s complicated” or “just trust the experts”. It’s a combination of structural defenses and collaborative tools that let you hold more of the world in view at once.
1. The Feeling of Being One-Shotted
There’s a meme about ayahuasca that resurfaces every few months:
This is what it is to be one-shotted: a single overwhelming perception rearranging a mind around itself.
But while the meme paints a picture of folly, the most insidious one-shots don’t look like this.
They look like rigor. Like strict adherence to the paradigm that sees the truth.
The trigger can be an event (a trip, a trauma, a spiraling LLM output) or a tool (a model that slowly becomes the only trusted lens).
Event-based one-shots happen in an instant: the before/after moment.
Tool-based one-shots creep in over years.
Both collapse the world’s dimensionality. Both mistake coherence for completeness.
Although event-based one-shots get all the attention, this essay focuses on tool-based one-shots, which are subtle, prestigious, and everywhere.
2. How Smart People Get One-Shotted
The usual story: people love simple explanations because they’re cognitively lazy.
But that’s not what I see. The people who get one-shotted hardest are often the most rigorous, the ones who’ve spent years mastering a domain.
Here’s the actual mechanism: experts don’t just find factors, they export causal structure.
What are factors?
A factor is any variable that helps explain why something happens. Interest rates affect housing prices. Childhood nutrition affects adult health. Lead exposure affected crime. Rule of law affects economic development.
Identifying such factors is good. It’s how progress happens. Domain experts isolate specific mechanisms and improve them, and our understanding of them. We need that. Civilization runs on specialization.
What can go wrong?
The economist doesn’t just notice that incentives matter in markets. Over time, every problem becomes an incentive problem.
The therapist doesn’t just see attachment wounds in relationships; suddenly attachment explains addiction, politics, art.
The engineer doesn’t just optimize systems; optimization becomes the lens for parenting, friendship, meaning.
The sociologist doesn’t just observe modern oppression; oppression is the unifying theory of history, government, and culture.
The problem is metastasizing the local frame: acting as if every problem can be solved like your field’s problems are solved, or assuming that the root causes in your discipline are the root causes of every other problem. The tool forgets it’s a tool.
3. The Self-Own
Yes, I know: writing this, I’m doing the thing. Foregrounding one factor (in this case, one-shottedness) and building an essay around it.
I’m guilty as charged. But the alternative is silence. To describe anything, one must spotlight one slice of reality.
The honest posture is humility about scope.
And since honesty requires self-disclosure, and skin in the game:
One-shottedness isn’t always wrong. Depth often beats breadth. Commitment can beat optionality.
Raising a factor isn’t claiming it’s the factor; it’s saying,
“this dimension deserves more weight in your model.”
That’s activism.
Ideology would be claiming this new lens explains everything.
4. Prioritization ≠ Ontology
The bureaucrat says: “I won’t pull any levers because my model says it’s complicated.”
The activist says: “I’m pulling this lever because it’s high-impact given my model.”
The ideologue says: “This lever is the only one that matters because my model is reality.”
Same factor. Different epistemic hygiene.
Ideologue
Every field begins as a bright insight about one slice of reality. Then the slice ossifies into identity. The insight becomes a tribe. The model forgets it’s a model.
When you meet a racist, a communist, a crypto-utopian, a technophobe, you’re meeting different flavors of the same cognitive pattern: one-shotted by a single factor. One cause, many evils. One cure, all goods.
Their rhetoric differs, but the structure’s identical: collapse a multidimensional world onto one axis, inflate that axis into morality, then demand obedience. It feels righteous because simplicity is a kind of power.
Bureaucrat
There’s a mirror danger on the other side: analysis paralysis. The person so aware of complexity they can’t act at all. They drown in dials, afraid to touch any without permission from every other part of the system. And they want to stop you, too.
Activist
The healthy stance is the activist: someone who sees many dials, chooses which to turn first, and remembers it’s a choice.
Not omniscient, not paralyzed. Just intentional.
5. Why This Persists
One-shottedness isn’t a bug; it’s an attractor sustained by at least five forces:
Cognitive demand
Multi-factor models are expensive. Every time you reason, you’re juggling variables, weighing tradeoffs, holding conditionals in memory. Single-factor models are cached. Under time pressure, you default to the simplest frame you have.
Learning curves
Mastering one domain takes a decade. Understanding how three domains interact? No institution teaches that. PhD programs train depth; nowhere trains synthesis. Even if you want multi-factor competence, where do you get the prereqs?
Social legibility
Institutions select for identifiable experts. “The incentives guy” gets hired, funded, booked. Someone who thinks “incentives AND norms AND path dependence” is illegible—harder to route, trust, coordinate with. One-shotted people rise faster because they’re fungible within their niche.
Narrative asymmetry
Big conversions make good stories. “I took ayahuasca and realized everything is trauma” (compelling). “I updated my model from 60% X to 55% X after reviewing meta-analyses” (no one shares that). The stories that spread are one-shot stories, even if most learning is incremental.
Coalitional pressure
Tribes punish hedging. If you’re “the communist” and admit markets work sometimes, you’re suspect. If you’re “the YIMBY” and say “it’s not all zoning,” you dilute the coalition. Purity signals commitment. Even if you privately hold multi-factor views, public hedging has costs.
These forces reinforce each other. Cognitive demand makes you cache one model. Social pressure makes that model your identity. Narrative selection spreads your conversion story. Coalitional dynamics lock you in.
This is one reason “just think harder” fails. You’re not fighting a bug. You’re fighting an attractor state with 5+ gravity wells.
6. Thinking Harder Is Now Easier
Another reason “just think harder” fails is that it’s hard. Historically, multi-factor reasoning meant years of study to prepare, only to be exhausted by every new question.
Now, nuance is almost free.
AI lets you ask, in seconds:
Collaboration with AI can take many forms:
The point isn’t to believe AI. It’s to use it as a lens—to see more than you could unaided. A single, perfectly coherent LLM output can createan event-based one-shot, a flash of false certainty. But using AI to map the landscape of disagreement is its antidote.
It’s imperfect, but fast enough to keep you from joining a crusade based on a cartoon.
Trust The Experts?
The antidote isn’t “trust the experts,” either. Experts disagree far more than most people realize: if they didn’t disagree, we wouldn’t need so many. Historically, nuanced consensus was accessible only by participation (joining the guild) or proclamation (waiting for a bureaucratic committee).
Now we can synthesize consensus instantly. AI can generate a snapshot of where experts align and where they don’t. You no longer have to trust one priest; you can triangulate the whole pantheon.
That makes complexity actionable. “It’s complicated” is no longer an excuse for paralysis. It’s a prompt to run the quick simulation before you act.
7. Structural Defenses
For now, AI tools are not enough. You also want structural commitments that make capture costly.
Diversification
Don’t let any single belief, role, or income stream define you.
If losing it would erase you, you’re already captured. This is why people should have hobbies. Why activists should have friendships outside the movement. Why you should be able to name three things you’d still care about if your current obsession turned out to be wrong.
Optionality
Preserve exit ramps. The person who quits their job the week after their ayahuasca trip burned their option to wait and see if the insight holds. The activist who makes their entire social world about one cause can’t pivot without losing everyone. Keep low-cost ways to test new models before committing. Keep relationships that don’t require ideological purity. Keep skills that transfer across worldviews.
Reversibility
Favor commitments you can test before they’re irreversible. The danger isn’t permanence itself; it’s the unexamined rush to make something permanent while under the spell of revelation.
Contextualization
Don’t force every speaker to issue caveats. That’s bureaucratic honesty theater. The responsibility lies with the listener. Before letting a single-factor claim rewrite your worldview, ask:
Zealotry spreads through listeners who mistake partial truth for total truth. Contextualization is inoculation.
Waiting
When clarity feels divine, wait ninety days. Real insights survive time; one-shots evaporate in contact with errands.
Coda
To notice a factor is good. To study it deeply is civilization. To export its grammar to all domains is delusion.
Every knob you turn moves others you can’t see. Single-factor stories make great openings and terrible endings.