Zero-Vector Design: You Will Move Planets
A new manifesto for how products and services are created in the era of AI
For thirty, forty, fifty years, design has been an act of translation. The moment digital products went from being built by the builders who coded them, to being split by design and development, that created a gap that only got wider decade after decade.
We draw pictures of things and hand them to other people who build fifty percent of the vision. We call this a process. We call this collaboration. We call this the way it has always been done.
It does not have to be this way.
I have spent almost 30 years in digital design and development watching the same pattern repeat. The tools changed every decade. Photoshop to Sketch to Figma. Waterfall to Agile to Lean. The labels shifted. The process never did. You think something, you represent it, you hand it off, someone else interprets your representation, they build their interpretation, you review it, you file tickets about the gaps between what you meant and what they built.
I am done with that. Every single one of those steps is lossy compression. Every handoff is a game of telephone. And if you have felt that friction, the translation tax, the “that is not what I meant,” the slow bleeding of intent across the pipeline, you might be done too.
This is Zero-Vector Design. Not a tool. Not a framework. A discipline. And the principles behind it did not arrive with AI. They have been building for sixty years.
The Lineage
In 1968, Douglas Engelbart sat in front of a camera in San Francisco and showed the world what a computer could be. A mouse. Hypertext. Collaborative editing. Real-time video conferencing. He called it the Mother of All Demos. It was not a product launch. It was a declaration of intent: the computer is an instrument for augmenting human thought.
Five years later, at Xerox PARC, Alan Kay and his team built the Alto, the machine that became the blueprint for every personal computer that followed. Windows. Icons. Menus. The graphical interface. Kay understood something profound: the tool shapes the thinking. Change the medium, change the mind.
Every generation since has tried to close the gap between what a person envisions and what actually gets built. Desktop publishing. The web. Agile. Design thinking. Lean. Each one shortened the distance. None of them eliminated it.
The translation layer survived every revolution. Until now.
(Read about it in my article “The Any Key” 👇)
The Demo (Yes, It Is Real)
I am building a software product called Fictioneer. It is a story development platform for fiction writers, think Notion meets Scrivener, but with AI that actually understands your characters, your plot, your world. Four major application containers. Twenty-nine database tables. Thirty-two migrations. A React frontend, a FastAPI backend, a RAG pipeline indexing over 400,000 words of my own published fiction.
I am building it with a crew of AI agents. Each one has a name, a personality, a domain. Heavy handles the frontend; warm, steady, “nothing to worry about.” Siddig owns the backend; methodical, precise, never uses contractions. Qin audits the code; sharp, clinical, does not mince words. I direct them like a film director runs a set.
This is not vibe coding. This is architecture. This is a real product, with real complexity, shipping real features to real users. And I am the only human on the team.
Everyone Focused on the Code
When I show people Fictioneer, they focus on the code. “Oh, AI can generate React now.” That is the wrong takeaway. Entirely the wrong takeaway.
The code is the least interesting part.
What people are missing, and I mean genuinely missing, is that what is happening with Fictioneer is not a coding story. It is a design story. It is a research story. It is a strategy story. The agents did not just write React components. They made architecture decisions. They interpreted design intent. They coordinated across a codebase with dozens of files and thousands of lines, maintaining consistency I could not achieve alone.
And here is the mindset shift that matters: the agents are not my assistants. They are my crew.
That distinction is everything. An assistant waits for instructions. Crew members understand the mission, anticipate needs, and act with judgment. When I work with AI agents now, I am not prompting a tool. I am directing a crew. They are not assistants. They are crew.
And that is a fundamentally different conversation than “AI writes code.”
See, the discourse right now has this massive blind spot. We are having a debate about whether AI will replace developers, and meanwhile the entire design pipeline, from first insight to shipped product, is quietly transforming underneath us. Research. Synthesis. Ideation. Prototyping. Validation. Build. Every single phase.
Not just the last one. Every single one.
The question is not whether AI can code. It can. We have established this. Let us move on. The question is: what happens when you remove every intermediary tool between your intention and the artifact your customer touches?
Zero-Vector: Intent to Execution
Let me define what I mean, because I am coining this term deliberately.
In physics, a zero vector is the point of pure potential. Magnitude without a predetermined direction, all possible directions simultaneously. It is the origin point before a force is applied. I am no physicist, but the analogy is precise: in traditional design workflows, your intent passes through a dozen translations before it reaches the customer:
You think something.
You sketch it.
You wireframe it.
You mock it up in Figma.
You annotate it.
You hand it off.
A developer interprets your annotations.
They build it.
You review it.
You file tickets about the gaps between what you meant and what they built.
Every handoff is a game of telephone. Zero-Vector Design eliminates the intermediate vectors. Here is the core formulation:
Your intent moves directly to the artifact. Not through a chain of translation tools, but through agents that understand context, make informed decisions, and produce the actual thing. Not a picture of the thing. Not a prototype of the thing. The thing.
No Figma. No Sketch. No Framer. No handoff document.
Before the pitchforks come out: I am not saying these tools are worthless. I am saying the requirement for them as intermediary steps is evaporating. That is a different claim, and it is an important distinction.
This is not about removing design thinking. It is about removing the translation tax we have been paying for decades. The thinking remains. The craft remains. The intention, frankly, becomes more important, because you are no longer buffered by three layers of interpretation between your idea and its execution.
(Read about this topic in this article 👇)
The Research Phase Did Not Disappear. It Accelerated.
Think about what this actually looks like across the pipeline, because I think people are not going hard enough on this question.
Start with research. Market research, competitive analysis, landscape scanning. The unglamorous foundation that everything else sits on.
Old way: Commission a report. Wait weeks. Read a fifty-page PDF that is already stale by the time it hits your desk. Maybe you do some of it yourself, burning days on desk research that feels productive but is really just organized reading.
New way: Point an agent at your competitive landscape and say, “Give me a structured analysis of how our top ten competitors are positioning their onboarding experience.” You get a structured, sourced analysis in minutes. And the crucial part: you can push back on it in real time. “Go deeper on the three using progressive disclosure.” “Compare this to the patterns in Tidwell’s framework.” “Now synthesize the three strongest approaches into principles I can use.”
That is not a report. That is a research conversation. And it compresses what used to be weeks into hours without losing depth. If anything, you gain depth, because you can actually iterate on the research instead of treating it as a one-shot deliverable.
Jobs-to-be-done interviews? Same transformation. You still do the interviews, you must still do the interviews, the human contact is irreplaceable, but the synthesis happens during the session, not in a debrief two days later when half the nuance has evaporated. Your agent is listening, structuring, pattern-matching across sessions in real time. You walk out of four interviews with a structured JTBD framework, not a stack of sticky notes and a vague sense of themes.
The Diamond Breathes Differently Now
The double diamond assumed certain constraints. It assumed that diverging required weeks of exploration. It assumed that converging required painful prioritization because you could only build a fraction of what you explored. It assumed that moving from one phase to the next required translation: from insight to concept, from concept to prototype, from prototype to code.
What happens when translation becomes optional?
The phases do not disappear. Discovery is still discovery. Definition is still definition. You still diverge, then converge, then diverge again. But the walls between the phases become permeable. You can move from an insight to a working prototype in the same afternoon. You can validate an idea with a functional artifact, not a paper mockup. You can iterate on the real thing instead of iterating on a representation of the thing.
Christensen called the underlying principle of jobs to be done “what is the customer hiring the product to do?” The progress a customer is trying to make in a given circumstance. The job of the double diamond was always to manage the cognitive load of design under constraints. But the constraints have changed. The diamond still works. It just breathes differently now.
In the old model, each phase had a hard boundary because the cost of transition was high. Moving from ideation to prototyping meant learning Figma, spending days on mockups, iterating on static screens. So you naturally wanted to be very sure of your direction before investing that effort.
When prototyping is nearly free, when you can say “build this” and have a working version in minutes, the cost of being wrong drops dramatically. And when the cost of being wrong drops, you can afford to explore more aggressively. Diverge wider. Converge faster. Test with real artifacts instead of proxies.
The diamond is not dead. It is turbocharged. Like how most of a rocket’s weight is the fuel it has to carry just to get to orbit, most of the labor of the process was the process itself. That can now be eliminated.
From Boulders to Planets
Let me be clear about something. These tools do not necessarily reduce your investment. They do not make the work easy. I want to be honest about this, because the “AI makes everything effortless” narrative is both wrong and dangerous.
What these tools do is change the scale of what your effort produces.
Think of it this way. Before, it took everything you had to push a boulder up a hill. All your skill. All your experience. All your craft. And at the top, you had a boulder on a hill. Meaningful. Real. But bounded by what one person, or one team, could physically push.
Now? It still takes everything you have. The effort is not less. But you are not pushing a boulder anymore. You are changing the orbit of a planet.
It still takes everything you have. The effort is not less. But you are not pushing a boulder anymore. You are changing the orbit of a planet.
Same energy. Same investment. Same craft and intention and hard-won expertise. But the scope of what that effort produces is fundamentally different. The force multiplier is not about doing less. It is about doing more with the same fire.
And honestly, this is what excites me most. Not that the work gets easier. That the work gets bigger. That the ambition you have always had but could never fully execute, the vision that always got compromised by resource constraints and timeline pressure and “we do not have budget for that” suddenly becomes possible.
Ideation Without the Translation Tax
Consider what changes in the middle of the pipeline.
Ideation. The creative phase where you are supposed to go wide, think divergently, generate possibilities. In practice, in most organizations, this phase is brutally constrained.
You brainstorm on sticky notes.
You sketch on whiteboards.
And then you have to translate those sketches into something a developer can build.
That translation can take weeks. And by the time it is done, the energy of the original idea has dissipated.
With agent-first workflows, ideation and prototyping collapse into the same motion. You describe what you are thinking. The agent builds it. Not a mockup, a working version. You react to the real thing. “No, the navigation should feel more like a conversation.” The agent adjusts. You iterate on the artifact, not on a picture of the artifact.
This is the part that should make every designer lean forward: you are finally designing the actual thing. Not a proxy. Not a specification. The thing itself. The gap between imagination and execution shrinks to nearly nothing.
And the skills that matter most in this world? Not your ability to use Figma (though that is fine). Your ability to think. To articulate intent. To know what good looks like. To recognize when something is not working and say why.
The tools change. The craft does not. A good craftsperson never blames their tools.
The Seven Principles
Zero-Vector Design is not vibes. It is doctrine. Seven principles, forged across decades of building things that ship. These are intentionally opinionated — they exist to help you make decisions in moments of indecision or crisis, not to be generic or all-purpose.
Principle Zero: Take from all that is around you and make of it something more.
That is the soul of it. Creation as synthesis. Everything you have learned, everything you have seen, everything you have failed at, it all becomes material. Now, the principles:
I. Work in the Medium. A chef does not draw a picture of a meal. Do not abstract yourself away from the thing you are actually making. Hands on the rock. No gloves.
II. Boundaryless by Nature. Zero-Vector prescribes no tool, no discipline, no lane. The boundaries between “designer” and “developer” and “researcher” were always artificial. When one person with AI agents can do all of it, the boundaries dissolve.
III. The Medium is the Message. McLuhan was right. The tool shapes the thinking. Choose tools that expand what is possible, not tools that calcify what you already know.
IV. Compound Your Leverage. Every system you build should make the next system easier. Every agent you train should make the next agent smarter. Stack your gains.
V. Intentional Impermanence. Nothing you build will last forever. Design for change. Build things that can evolve, be replaced, be transcended. The artifact is not the point. The capability is the point.
VI. Living Systems Over Static Artifacts. A design system that cannot evolve is already dead. Build things that breathe, that respond, that grow.
VII. Authentic Voice. AI can generate text. It cannot generate your voice. Preserve what makes you irreplaceable. The craft is in knowing what to keep and what to delegate.
The Scaffold: Investiture
Principles are nice. But you need somewhere to start.
Investiture is the Zero-Vector starter scaffold. It is an architecture that teaches your AI to write clean code, folders, separation of concerns, guardrails. A spaghetti-free zone for building with AI agents.
Clone it. Open Claude Code. Start building.
It is free. It is open source. It is named for the fundamental magical energy of Brandon Sanderson’s Cosmere universe, the force that makes thought manifest. Which is exactly what it does: infuse your intent with Investiture.
The scaffold handles architecture so you can focus on creation. Four layers: design system, content layer, core logic, services. Each one has a job. None of them leak into each other.
You do not have to become an engineer. You just have to build inside structure that already knows the rules.
See Investiture here: https://zerovector.design/investiture
The Orbit Awaits
I started by building a product I could not have built alone. Twenty-nine database tables. Four application containers. A RAG pipeline over 400,000 words of my own fiction. A crew of AI agents, each with a name and a role and a voice. Fictioneer is not a weekend project. It is a real product, with real architecture, solving real problems for real users.
And I am the only human on the team.
But the real story was never about Fictioneer. It was about what building it represents: a fundamental shift in the relationship between intention and execution. Between the designer and the design. Between the sixty-year dream of closing the gap and the moment the gap actually closes.
Zero-Vector Design is not a rejection of craft. It is the fullest expression of craft, where your thinking, your taste, your hard-won understanding of what makes things work for people moves directly into the world without losing fidelity along the way.
The entire pipeline transforms. Not just the code. Not just the flashy demo. Research. Synthesis. Ideation. Prototyping. Validation. Build. Every phase. Every handoff. Every moment where intent used to degrade through translation.
This is not a framework launch. It is a movement. And the only price of admission is the willingness to start.
“If you want to build a ship, don’t drum up people to collect wood and don’t assign them tasks and work, but rather teach them to long for the endless immensity of the sea.”
