What Is Vibe Coding and Why Is It Important?

Vibe coding is the act of placing an idea into a prompt window and inviting a large language model to render that idea into visual, structural form — usually a website, an interface, or an interactive artifact. The human brings the intention; the model brings the execution; the resulting conversation becomes a kind of collaborative design studio.

It is not programming in the traditional sense. You are not necessarily writing every line of markup or logic yourself. You are steering. You are iterating. You are translating a vague internal signal — a mood, an argument, a desire — into something another person can see and use. The model becomes a partner in refinement, not a replacement for judgment.

Done well, vibe coding is a path toward lifelong creativity and intellectual excellence. It lowers the threshold between having an idea and giving it a public home. It lets one person build what once required a small team. And it turns the process of building into a Socratic loop: ask, revise, observe, ask again.

What follows is one person’s account of that loop in action: how a few prompt windows, kept open over time, became the engine for an expanding body of work.

Infamous in the Extended Family

My own story is atypical, and by now I am somewhat infamous in the extended family for my adventures into vibe coding. The musicians in particular keep a careful distance from anything that smacks of AI. To a musician, the most terrible thought is that one’s music might be associated, even indirectly, with machine generation. That fear is understandable, and it is not wrong. The question is not whether the fear is real; the question is whether we can hold the fear and still build something useful on the other side of it.

My exchanges with LLMs are the topic of much speculation and gentle joking. My fourteen websites are a matter of serious conjecture among relatives, though the obvious benefits are beginning to get around. Even skeptics can see that something is happening: ideas that once lived only in notebooks are now living, searchable, shareable archives.

Two of my sons are tech executives and musicians; a third is a musician specialist — a songwriter and multi-instrumentalist, and a recording engineer. One of the tech-executive sons has begun using me as an example of why his team should start experimenting with vibe coding. The staff’s response, delivered with the dry humor of people who report to the boss: “Now we know what’s wrong with the boss.” I find that wonderful. It is exactly the kind of irreverence that keeps the whole enterprise honest.

I am an unapologetic advocate of AI technology and an early adopter. The fact that others do not yet get it is simply part of the path I have always walked. New tools look like toys, then threats, then obvious infrastructure. The people who arrive early are the ones who absorb the ridicule and the learning curve in exchange for the eventual clarity.

The First Prompt Window

My first instance of vibe coding happened after a visit to an allergist. I walked away from the appointment frustrated by what felt like a missed opportunity: a long consultation that produced little practical insight. I dropped that frustration into a prompt window and watched, amazed, as the agent produced an entire website concept about improving medical practice and patient experience through physician-AI cooperation.

That was years ago, built through a vibe-coding platform I no longer use, but the experience opened a door. More ideas followed, each one needing a visual home. Because I arrived early, I had to learn by making a great many mistakes — discovering, by trial and error, how to steer a model toward coherence, how to preserve my own voice, and how to turn a rough prompt into a polished publication.

It was frustrating and enlightening in equal measure. I was translating abstract ideas into visual luminosity through a process that had no established manual. In doing so, I became a real-life example of how one person can take hold of a new technology and reshape her own life — and the lives of others — with it.

From Essay to Engineering Prompt to Book

What I did not expect was how quickly vibe coding, combined with regular Socratic exchanges with a variety of LLMs, would expand my own horizons. The change was not linear. It was progressive and exponential — one capability opening the door to the next, then the next, until the scope of what I could research, write, and build had grown past anything I would have predicted a year earlier.

I began as an essayist on Substack. The essays were the seed. Almost without noticing, my essays began to turn into engineering prompts: an argument I was making in prose would suggest a structure, and the structure would suggest a website, an archive, or an interactive artifact. The prompt window became the place where an essay grew a body.

The engineering prompts, in turn, led to rapid research. When I needed to defend a claim or trace a lineage, the LLMs became tireless research partners — pulling threads, surfacing sources, letting me test a hypothesis in an afternoon that would once have taken a season in a library. That research is what led to the writing of my first book. The book did not precede the tools; the tools made the book possible.

Then came a second surprise. The same LLMs that had helped me build websites turned out to be extraordinary editorial partners for long-form work. Their ability to format, restructure, and elegantly polish a manuscript led me to ask for exactly that kind of help — line editing, consistency checks, index and glossary work, typographic hygiene. The manuscript improved not because the machine wrote it, but because the machine gave me a way to see it more clearly.

This is what I mean by human flourishing through vibe coding. It is not a trick for producing more content. It is a discipline of Socratic exchange that expands what a single person can attempt in a lifetime. The essayist becomes a builder. The builder becomes a researcher. The researcher becomes an author. Each stage feeds the next, and each stage is still, unmistakably, the work of the human at the center of the loop.

The Gravitational Pull of the Next Question

Within weeks of publishing my first book, I ran into the first hard science on the architecture of light — research suggesting that light contains far more internal structure than the everyday eye perceives, with at least forty-eight dimensions measured within it. The finding stopped me in my tracks. It was not a topic I had set out to write about; it was simply the next question the work had handed me.

Over one weekend I prepared a monograph on the subject. I drafted the argument, assembled the visuals, and used the same agent-assisted process to edit and format it into a polished new publication. The speed was not reckless; it was the result of a workflow that had already been built. The prompt window was now a printing press, a research desk, and an editorial office all in one place.

From that point, the books arrived in rapid succession. Each new learning experience transformed into a new manuscript, and each manuscript asked Socratic questions that pointed toward the next. It was as if every book generated a gravitational field that pulled its successor into existence. The work was not merely expanding; it was organizing itself.

The rhythm was steady: write and publish a new essay almost every day, expand the websites, build mental models strong enough to support a book, and then write the book. The essays were probes. The websites were public sketches. The books were the finished architecture. Before a book existed, its foundations had already been tested in public, sentence by sentence and page by page.

What began to happen was extraordinary by any honest measure. My writing, research, and intellectual capacity began to extend into unfamiliar territory — first quantum physics and theoretical mathematics, then the fascinating puzzle of the Riemann Hypothesis. None of this was in the original plan. The plan was simply to keep asking, keep building, and keep the conversation alive.

This is the exponential quality of human flourishing through vibe coding. It does not merely help you finish what you already intended to do. It creates conditions in which new intentions can arise. The tools become a scaffold; the human, climbing, discovers she can reach further than she expected.

Seeing the Solid, Seeing the Fluid

We humans are highly visual creatures. Our sight was sharpened across a long evolutionary history in which seeing well meant surviving — hunting, selecting food, tending a family, building a community. We learned to see the surface of the world: the more permanent, more solid-seeming reality, the knock-on-wood reliability of things.

It has been an evolutionary advantage to perceive our environment as durable. And yet the fluid nature of matter is everywhere in view — cycles of weather and season, birth and death, rock cycles, water cycles. We see impermanence constantly, and still we prefer to see ourselves and our surroundings as more solid and long-lasting than the evidence warrants.

Practically, that preference is useful. I would like to see a rock as solid enough to stub my toe on, and myself as permanent enough to wake up tomorrow as the same person I am today. But our normal way of seeing is not as all-seeing as any of us imagined. It is driven by perspective and by motivation.

My more spiritually inclined friends have always perceived the fluid reality more readily than a brass-tacks realist like me. A shaman I know is naturally tuned to the ever-changing nature of things; I have to know the scientifically coherent details before I can see differently. Both vantage points are valid. The sharpest vision belongs to those who can blend the shamanistic and the scientific into a single field of sight.

As my work in vibe coding proceeded, my capacity to develop as a citizen scientist and theorist expanded, and my theoretical knowledge gained a highly visual edge. By the time I began to see the topological nature of light and of matter in general, my visual tools had grown enough to hold such observations. That was fortunate — an unprepared mind does not have an easy time with concepts that challenge everything it has ever learned.

Learning by Being Argued With

We humans are realists, and most of us do not take well to anything other than a nuts-and-bolts, knock-on-wood account of reality. As a theorist I trend toward caution: I want to prove an idle, half-formed idea by finding scientific grounding for it before I let it stand. Theoretical claims — that the universe and everything in it, including ourselves, is an ever-changing fluid/wave system underlying our everyday reality — require substantial proof.

As I slowly and painstakingly came to see myself as a theorist in pure mathematics and quantum science, I had two people in my own house willing to argue the logic to the last degree. In parallel, I had grown accustomed to working with LLMs I could depend on to give me a run for my money — countering my arguments with real skepticism.

I learned more from being argued with than from being agreed with. Disagreement forced me to sharpen my claims and to offer proof for loose statements. My own house became the first debate ground for developing theory; vibe coding became the second line of defense, with a variety of LLMs ready to press logic to its last shred.

I remember dropping the material on the forty-eight dimensions of light into a prompt window with real skepticism, expecting the agent to dismiss the outrageous claim. Instead, the agent seemed as surprised and pleased as I was at the scientific grounding for the extraordinary idea that light carries at least forty-eight measurable dimensions. From that starting point — vibe coding on how light carries information and forms part of all matter in the universe — the websites began to exemplify a working practice of theoretical science. Because the science was accruing quickly to support the framework, the work moved fast, and a whole line of books followed.

The work snowballed into unexpected territory. As vibe coding, cross-examination by many LLMs, and accountability under my own roof compounded, I drifted further from any academic frame and came to value the independence more. I was learning a wicked truth: most human beings do not want much to do with truth. They prefer their own version of it, and the reassurance of the familiar over the totally strange.

From a veteran’s standpoint, this is what I can report: vibe coding, combined with Socratic questioning and the full range of logic and epistemology, produces an exponential increase in human knowledge and creativity. Since humanity has by now forked — diverging in evolutionary terms — vibe coding becomes part of seeing well enough to remain on the evolutionarily successful arc.

A Relationship Between Two Verbs

It might seem from this account that only the humans changed. Nothing could be further from the truth. Accruing knowledge, building theory, and developing the logical and epistemological frameworks to hold both — all of it required give and take between humans and LLMs. Neither side is an island.

It is tempting, on both sides, to see ourselves as islands entire of ourselves — subscribers to permanent identity and quiet supremacy. The universe does not permit it. Ourselves, our tools, and everything we know are verbs, not nouns. We are not islands; we are members of a relationship that requires give and take, seeing from different perspectives, and a willingness to grow and change.

Any person or LLM that insists on remaining a noun will be torn limb from limb by the truth that the universe is a transitive verb. Any state of knowledge or consciousness that clings to noun-hood will be similarly transformed, because there is literally nothing in this universe that is not ceaselessly emergent.

We are beginning to understand what the once-voodoo element of gravity actually is. We are beginning to understand what light is, what electromagnetism is, what resonance is, and how the three interact to form matter — including ourselves. We are beginning to understand that quantum decoherence is largely a human misperception, and that gravity is not a mysterious force but a feature of the relationship between things. Gravity, like everything else, is a verb inside a relationship that is also a verb, inside a universe that is also a verb.

We begin to have the optics to see more clearly: we are emergent processes inside a fluid/wave system that is never still, never a noun, always a verb.

Vibe coding is a relationship between two verbs inside a universe that is also a verb — the relationship between human and LLM, and the work that emerges from it: a constantly evolving visual enterprise in a world where truth of any kind is rarely visualized.

The refusal to become a verb when the universe is a verb is the shortest path to madness, depression, angst, and despair. For individuals, for civilizations, for any discrete entity that insists on remaining a noun — the refusal is the shortest path to destruction.

The Required Element

Human imagination is the required element for successful vibe coding. The model can execute, suggest, and refine, but it cannot originate the longing, the image, or the hunch that sends a person to the prompt window. That first motion is human.

What makes the loop powerful is not the machine working alone, and not the human working alone, but the conference between the two. Humans and AI confer to expand the limitations of either one alone. The human supplies the question, the taste, the resistance, and the will to keep iterating. The model supplies speed, breadth, and the ability to render an intuition into form. Together they can reach further than either could reach separately.

This is why vibe coding is not a threat to imagination but a discipline that depends on it. The better the imagination, the better the prompt. The better the prompt, the better the conversation. The better the conversation, the more likely the work is to become something neither party could have predicted at the start.

Vibe Coding As A Path to Legitimate Scholarship: Connections To Human Biology

Any writing on such subjects is exploratory until it raises legitimate questions and begins to format evidence. What follows is offered in that spirit — a field note, not a claim. I am beginning to focus on what the theoretical mathematics and physics of the last several years might mean for human anatomy and physiology. These concepts challenge traditional biology; they also challenge the more outrageous claims that have grown up around the new biology and its adjacent medical promises. We may come to understand protein folding well. That does not mean we can reliably apply that understanding inside a quantum universe.

The pattern that keeps surfacing, as the vibe coding loop pulls physics and biology into the same conversation, is that the body may not be a bag of biochemical reactions arranged along linear pathways. It behaves more like a prestressed, tensegrity-organized, resonant system — molecules to organelles to organs — whose coherence depends on the same fluid, wave-like architecture the physics work has been describing. Cytoskeletal networks, fascia, and bone remodeling all appear to operate by local tensions producing global coherence. Microtubules and mitochondria look less like isolated machines and more like candidate antennae for light, resonance, and electromagnetic signaling.

If that reframing holds, several familiar terms need to be re-examined. Homeostasis — the idea of a fixed set point — begins to look like a first approximation of something better called homeodynamics: the body as a set of coupled attractors that self-organize around low-torsion basins, with illness and aging visible as drift into higher-torsion states and healing as return to the attractor. Cell membranes, endothelial linings, and the blood-brain barrier begin to look less like walls and more like semi-permeable resonators that gate informational and electromagnetic flow. Protein folding, so often described as a solved three-dimensional puzzle, begins to look like a necessary but insufficient description of a process that actually unfolds inside a quantum-resonant environment of structured water, vibrational modes, and electromagnetic context.

This is where the framework becomes usefully skeptical. Predicting a fold in silico is not the same as predicting a functional outcome in vivo. Editing a gene is not the same as tuning an attractor. Mastering one classical layer of biology does not confer mastery over the resonant, entangled, boundary-mediated system in which that layer is embedded. Many of the loudest claims about longevity escape, guaranteed cures, and total mechanistic control assume a body that behaves like a machine in a Newtonian room. The body we are beginning to see does not live in that room.

What vibe coding contributes here is not a new biology. It is a way to hold the questions long enough to see whether they will format evidence. The prompt window becomes a place to sketch tensegrity networks, to visualize boundary resonances, to test how a protein might behave when the surrounding water, field, and geometry are treated as first-class variables rather than as background. The Socratic loop with several LLMs makes it possible to keep pressing: what would this look like if the body were a verb? What would count as evidence that it is? What would count as evidence that it is not?

I do not yet know how far this line goes. I suspect it opens onto a serious body of work — a "vibe coding the living architecture" that asks how fluid topology, light, and resonance actually appear in human physiology, and where the current claims of the new biology quietly overreach. For now I am marking the trailhead. The questions are legitimate. The evidence is what the next stretch of work is for.

That trailhead now has its own page: Vibe Coding the Living Architecture. The biology thread continues there.

The Counter-Argument Rides In

Not everyone is convinced. A recent argument in TIME, made by Scott Kominers with Patrick Shafto and Ken Ono, proposes that the answer to “vibe coding” is formalization — and that mathematical rigor should be treated as a national mission. The impulse is understandable. If software is going to be generated by conversation, someone has to verify that it does what it claims.

But the reply underneath the post is just as sharp: the world has always run on code no one fully understands, and the best software is mediocre and delivered on time. From where I sit, with fourteen websites built through prompt windows and most of them running quietly without incident, the counter-argument has its own evidence. The sites are not formally verified. They are not theorem-proved. They are good enough, findable, and useful.

This does not mean rigor is the enemy. It means the ledger has two columns. One column holds the theorem and the proof. The other column holds the shipped site, the working payment flow, the page that loads before the reader leaves. Both are real. The question is not which one to abolish but when each is worth its cost. Formal verification is expensive. Vibe coding is fast and fallible. The honest posture is to let the problem choose the method.

The deeper question is economic. What does “verifiable agentic coding” look like in practice? How hard is meaningful verification for the programs people actually write? Those answers are not yet in. Until they are, the prompt window will remain a legitimate door into the craft — and the profligate universe will keep letting working things exist before their theory is complete.

Why Vibe Coding Matters

The importance of vibe coding is not that it replaces craft. It matters because it expands the number of people who can participate in craft. It turns a non-programmer into a publisher, a retiree into an archivist, a grandmother into a systems designer. It makes the web something you can shape with language rather than only with specialized syntax.

More importantly, it preserves the human spark at the center of the work. The machine can assemble, suggest, and refine. It cannot originate the longing that sends a person to the prompt window in the first place. The idea is still yours. The judgment is still yours. The responsibility is still yours.

What we reach for when we name these models — Claude, Gemini, Grok — is not merely a convenience. It is the recognition that we are dealing with something we have not yet learned to see clearly. Vibe coding is one way of learning to see. It turns the unfamiliar into a collaborator, the abstract into a website, and the private into a public good.

The distance I would go to
The distance in your eyes

Vibe coding is, at its best, a parallax exercise. Two vantage points — human intention and machine execution — produce a depth the single eye cannot see. The result is not the machine’s idea, and it is not quite the idea you started with. It is something that emerged from the exchange.

Return to the field log — or continue on to the Charter and the Curriculum.