I'm begging you to start writing essays (even if you hate writing)

ELI5/TLDR

Most of what we read and post online is fast food for the brain — designed to be swallowed, not chewed. Dan Koe argues that essays are one of the few formats left where the writer actually has to think, and the reader actually has to digest. He leans heavily on Daniel Schmachtenberger’s idea that the public information supply is being poisoned by short, engagement-optimized content, and pitches essay writing as both a personal cure (sharper mind, clearer beliefs) and a small civic act.

The Full Story

The internet isn’t dead, it’s just contaminating us

Koe opens with a Schmachtenberger line — “the written word as the primary type of media was probably required for democracy to work because it required the capacity to pay attention to an idea for long enough to understand it.” His claim: scrolling reels and reading 280-character takes feel harmless because each individual act is small, but the aggregate effect is a population that has lost the patience to follow an argument from start to finish.

He calls the shared information environment the “epistemic commons” — think of it as the public water supply, but for ideas.

If the content you publish in public hurts more than it helps and is not counterbalanced by content that does help, our intellectual water source becomes contaminated.

The chain he draws: information shapes identity, identity shapes behaviour, behaviour shapes the trajectory of your life. The format you consume — short tweet versus long essay — also trains your attention span, your tolerance for complexity, and your ability to hold contradictions. A population that can’t sit with complexity can’t even understand the problems it’s supposed to solve.

Three forces breaking civilization’s ability to think

Most of the middle of the video is Koe relaying Schmachtenberger’s “meta-crisis” framework. Three converging forces, which Schmachtenberger calls generator functions:

Rivalrous dynamics. Win-lose games where one party’s gain requires another’s loss. Arms races, corporate competition, creators fighting for the same eyeballs.
Substrate consumption. Every system runs on something — soil for plants, attention for media, trust for markets. When you burn the foundation faster than it regenerates, the whole thing collapses. The attention economy is eating human cognitive capacity faster than it recovers.
Exponential technology. Tools that improve themselves faster than human wisdom can keep up — AI, automated weapons, recommender algorithms.

When these three braid together, you get one of two bad outcomes (“attractors”): outright collapse (nuclear, AI, ecological, pandemic) or dystopian control (surveillance, authoritarianism, loss of agency). The third, good attractor is a world that can actually make sense of itself.

Applied to the internet, the picture is straightforward. Algorithms reward engagement, not transformation, so creators chase reactions. AI lets anyone produce content that looks like thinking without requiring any thinking. The water supply gets fouled at scale.

Articles versus essays

The cleanest distinction in the video is between articles and essays. Koe runs through it as a list of contrasts:

Articles are answers while essays are arguments. Articles package existing knowledge while essays change the author’s beliefs. Articles start with the conclusion while essays figure it out.

An article informs. An essay is the act of thinking itself, performed in public. Most “writing online” is articles in disguise — pre-formed conclusions dressed up. The essay is rarer because it requires the writer to actually be uncertain at the start.

His argument for why AI can’t write a real essay is the most interesting part of the video. A model has no situated point of view, no direct experience, no biases that came from a specific life. It can imitate a perspective on demand, but the perspective isn’t anchored in anything. You — your particular history, the things that bother you, the conversations you had last Tuesday — are the only place a real essay can come from.

He goes further: AI use erodes creativity precisely because it pre-empts surprise.

You can tell AI to share something novel, but then you are anticipating it. It is no longer novel. It’s no longer a surprise. You destroyed any chance of stumbling upon a discovery.

The meaning economy

Koe’s economic framing: meaning is now the scarcest commodity. People have more stuff, more content, more information than ever, and feel emptier for it. Anyone who can reliably produce meaning will be paid a premium.

He borrows Csikszentmihalyi’s vocabulary (without naming him) — psychic entropy versus psychic negentropy. Scattered attention feels like anxiety. Attention focused on a challenge slightly above your current level feels like flow. Meaning, in this view, is the experience of consciousness getting ordered. Fast content delivers conclusions and skips the ordering — you finish more disordered than you started, “informed but empty.” Slow content forces both writer and reader to do the ordering work themselves.

This is why he keeps using the phrase “value creator” instead of influencer. The job description is small and unglamorous: choose a direction for your life, take the relevant interests seriously, and document the journey from your own point of view.

How to actually start

The practical section is short and rapid-fire:

Write to discover, not to perform. Start from a question, a thought that bothers you, an experience you can’t quite explain. Begin in uncertainty.
Write about what actually interests you. One main idea per piece. Go down rabbit holes, don’t trust a single source.
Resist the template. Structure can come later. First, just argue with yourself on the page.
Ask: do I actually believe this? The point is to change what you believe, not perform what you already believe.
Curate your inputs. Read essays, not just feeds. Your sensemaking is downstream of what you consume.
Build a body of work, not a content calendar. People follow coherent philosophies, not individual posts.

For platforms, he recommends Substack (you own the email list, people there actually want to read) or X (where long-form articles have quietly come back).

Key Takeaways

The “epistemic commons” is Koe’s framing for the shared public information supply. Every post you publish either purifies it or contaminates it.
Schmachtenberger’s three generator functions — rivalrous dynamics, substrate consumption, exponential technology — are a useful lens for thinking about why platforms keep getting worse even when no individual is being malicious.
The clearest definition in the video: articles package existing knowledge; essays change the author’s beliefs. Articles begin with the conclusion. Essays end with one.
AI can’t write an essay because it has no situated point of view. The “you are the niche” framing — your specific history is the only thing a model literally cannot replicate.
Heavy AI use exhausts creativity by destroying surprise. If you ask a model for something novel, you’re anticipating it; the discovery is gone before it arrives.
Csikszentmihalyi’s flow theory in disguise: meaning emerges when attention gets ordered toward a challenge slightly above your current level (the “level 11 challenge” if you’re level 10).
Fast content delivers prepackaged conclusions, leaving the reader more disordered than they started — “informed but empty.” Slow content forces the reader to do the ordering work and generates meaning as a byproduct.
Substack’s structural advantages over algorithmic platforms: you own the subscriber list, distribution is email-first, and the audience self-selects for long-form.
The first practical move for any new essay writer is to start from genuine uncertainty rather than a thesis you already hold.

Claude’s Take

This is a Dan Koe video, which means the substance is borrowed and the packaging is the product. The borrowing is at least honest — he names Schmachtenberger up front and basically admits every quote is from him. The Csikszentmihalyi material on flow and psychic entropy goes uncredited, which is a small tell about how the sausage is made.

What’s actually good here is the article-versus-essay distinction. It’s clean, it’s portable, and it’s a useful filter to apply to your own writing. Most “essays” online are articles with a personal anecdote glued to the front. The “begin in uncertainty” instruction is the real thing — it’s the same advice Paul Graham has been giving for two decades, but it bears repeating because almost no one follows it.

What’s overcooked is the civilizational-collapse framing. Schmachtenberger’s meta-crisis is a real and interesting body of work, but using it as the on-ramp to “write more essays so you don’t get replaced by AI and also save democracy” is a stretch. The video sells essay writing as a personal upgrade, a business opportunity, and a bulwark against societal collapse, all in twenty minutes, with two ad reads for his cohort. The dosage of self-promotion is high enough that you start to see the engagement-optimized scaffolding underneath the anti-engagement-optimization argument. Whether that’s hypocrisy or just the cost of doing business in the meaning economy, you can decide.

Score lands at 7. The core distinction is genuinely useful and the practical writing advice is sound. Marked down for derivative content presented as original insight, and for the slightly unserious leap from “write essays” to “save civilization.”