Why I ran six versions of my visual brand before picking one.

Stop at version one and you get a generic personal brand

The reason most new personal brands look the same is that everyone stops on the first version that looks decent. Dark mode is not the problem. A cyan accent is not the problem. I use both, and so does half the AI personal brand internet. The problem is shipping just those defaults with nothing specific layered on top. No archetype. No vocabulary. No reason for the choices. The visual brand looks fine. The owner ships it. Six months later it reads as another stock dark-mode AI personal brand because no specificity ever got built on the surface.

I almost shipped version one. Then version two. Then version three. All three looked good. None of them felt like me.

The version that finally landed was version six. And version six was not version one with extra steps. It was a different visual brand entirely. Each iteration had moved the system somewhere new, and the final version was where every move I had tested separately came together.

If you are picking your own visual brand and the first version looks “good,” that is the warning sign, not the green light. Good is the trap. Specific is the goal. And specific takes more iterations than you think.

What you need before you start iterating

Before you run a version of any visual brand, you need three things on the page in front of the AI. A short version of your career. A short version of your audience. A short version of your positioning. Without these, every version comes out generic because the AI is filling in defaults.

I had all three because I had run a foundation marathon that morning. The career truth dump, the audience definition, the positioning line, the methodology. All of it was already in the chat. The visual identity was the last big move of the day, and it sat on top of work I had already done. If you skip the foundation and jump to visuals, you get a logo with no brand under it. The visual is supposed to be the surface of something specific, not the whole thing.

I have written about the one-day foundation marathon and what it produces. The visual brand is the last piece, not the first.

Versions one through three: the safe versions that almost shipped

The first three versions all looked decent. None of them were right.

Version 1: Clean Tech Noir

Dark backgrounds, electric cyan accent, Space Grotesk, DM Sans, JetBrains Mono. A clean grid layout. A nice hero. Some sample content blocks.

The whole thing looked like the default first answer the AI gives you when you say “AI personal brand, dark mode.” The hex codes were thoughtful. The typography was clean. The thing was forgettable. I sat with it and felt nothing. The first move was easy: cyan on black. The problem was that cyan on black with nothing else on top is what every AI personal brand is shipping. The system needed a reason to be those specific colors. V1 did not have one yet.

I told Claude this version was too generic and asked for the next one to push toward something specific to my career.

Version 2: The Architect (early version)

This is where the architect vocabulary first showed up. Blueprint grid background. Brand vocabulary words rendered as little tags across the hero. Outfit replaced Space Grotesk. The hero had a build-progress checklist. The architecture word started doing real work.

It still looked like a developer’s portfolio. The system felt like a Linear product page or a SaaS startup landing. Not a personal brand. The architect language was right. The execution was wrong. I told Claude V2 was more interesting than V1 but still felt like a software product.

Version 3: Jobs and Musk influence

Stripped back to Apple-level simplicity. A browser window mockup framing the hero. Ultra-thin (200) weight contrasted with bold (700). More breathing room everywhere. The aesthetic shifted from “engineer’s workstation” to “visionary’s studio.” Very clean. Very safe.

It looked like a default Notion site with better typography. I would not have remembered it the next day. The minimalism was so heavy that the personality leaked out. There was nothing to grab onto. Three versions in. Three versions wrong. None of them were broken. None of them were mine.

Versions four through six: the breakthroughs that built The Architect

Versions four, five, and six were where the visual brand actually showed up. Each one introduced a new element that survived into the final version.

Version 4: The Drafting Table

Full commitment to architectural blueprint aesthetic. Instrument Serif italic added as the display font. An engineering drawing title block at the top with project name, sheet number, version, and date. An isometric 3D system topology diagram showing all my businesses orbiting one center node. Material specification tables. A brand lexicon rendered as a numbered technical catalog. Amber added as the fifth color for warmth.

This was the most distinctive version yet. It also felt too academic for a personal brand. The pages read like a design review document, not a place a reader would land and feel something. The details were cold.

But the soul of the visual brand showed up here for the first time. The vocabulary, the typography, the spec-sheet rhythm. Half the final brand lives in V4.

Version 5: The Briefing

Top Gun energy injected. A fixed HUD status bar at the top of the page. Barlow Condensed 800 as the new display font. Athletic, condensed, built for speed. A mission briefing dossier format. “Call sign, doctrine, payload, target” rows replacing the spec rows from V4. Radar diagrams in thumbnails. Cinematic energy.

It was too military for a personal brand on its own. A mission briefing dossier reads great as a hero, less great as an ongoing publishing format. I would have felt like I was cosplaying my Air Force years.

But the cinematic punch worked. Barlow Condensed 800 was the right display font. The HUD bar was the right framing device. The energy was the right energy. Half the final brand lives in V5.

I told Claude V4 had the soul and V5 had the energy. Could the next version combine the architect vocabulary from V4 with the cinematic punch from V5, plus something warmer to keep it human.

Version 6: The Architect, final form

V6 was the synthesis.

Military precision from V5. Artistic soul from V4. Cinematic energy from V5. Human warmth (new). Five colors locked: Deck (dark base), Panel (slightly lifted background), Cyan (technical accent), Amber (human accent), Ink (text). Four fonts locked: Barlow Condensed 800 for headlines, Instrument Serif italic for emphasis, Inter Tight 300 for body, JetBrains Mono for spec rows and code.

A Method section showing three phases visually. A Flight Path origin timeline with career nodes. A Human section with amber-accented warmth (family, fuel, heroes, sanctuary). A Philosophy section with four bold-opinion cards. A closing signature.

V6 was not a version. V6 was the first thing that felt like me. It read like a field manual designed by an artist. Technical without being sterile. Cinematic without being cosplay. Personal without being soft. The architect’s vocabulary, the military precision, the cinematic punch, and the human warmth all lived in the same system without canceling each other out.

“Felt like me” was not about liking the colors. It was about the system matching my career, my taste, and the feeling I wanted readers to have when they landed on the page. If you do not have a filter that specific, every version is going to feel “fine.”

What six iterations actually buy you

Three things.

The first thing is permission to be wrong. Each version is a draft, not a commitment. If V1 has to look like the answer, V1 is going to be safe. If V1 is just the first draft on the way to V6, it can be wrong and that is fine. Iteration changes the relationship to the work.

The second thing is range. V1 explored generic. V2 explored architecture. V3 explored minimalism. V4 explored academic. V5 explored cinematic. V6 was the synthesis. Without the range, V6 would have been V1 with a few tweaks. The range is what made the synthesis possible.

The third thing is a lesson per version. V1 taught me what generic feels like in this category. V2 taught me where developer aesthetic stops working for a personal brand. V3 taught me where minimalism becomes invisible. V4 taught me what specific feels like, and the cost of going too cold. V5 taught me what cinematic feels like, and the cost of going too narrow. V6 was where every lesson got applied at once.

If I had stopped at V3, the personal brand would have looked like everyone else. If I had stopped at V4, it would have been too academic. If I had stopped at V5, it would have been too military. The number of iterations was not the magic. The discipline of treating each version as a learning step was the magic.

For me, six was the number that landed. For you it might land at four, five, or seven. The point is not to chase a number. The point is to keep going until something feels like you, not until something looks fine.

Common ways the iteration goes wrong

Three traps to watch for if you run this yourself.

Stopping early because the version “looks decent.” This is the most common one and the whole reason the post exists. If the version is decent, it is generic. Push it.

Treating each version as the answer. If V1 has to be right, V1 is going to be defensive. Brief the AI on a specific hypothesis to test (“more architectural vocabulary,” “more cinematic,” “more minimal”), and treat the version as a test result, not a finalist. The synthesis comes from the range, not from any one version.

Letting the AI lead the synthesis. The AI will keep generating new versions forever if you let it. The role of the human is to read each version, name what is working, name what is not, and tell the AI what to test next. If you just say “give me another one,” you get noise. If you say “more spec-sheet rhythm, less HUD bar, keep amber as warmth,” you get signal.

The iteration is a conversation, not a slot machine.

Put This Into Practice

If you are about to start the visual brand for your own personal brand, paste this into Claude or ChatGPT after you have already done the foundation work (career truth, audience, positioning, methodology). Without those four pieces, the AI fills in defaults and every version comes out generic.

The point is to keep going until something feels like you, not until something looks fine. The synthesis comes from the range. The range only happens when each version is allowed to be wrong on its way to right.

Why the iteration is the whole game

The visual brand is not picked. It is iterated into. The first version is research. The second version is reaction. The third version is calibration. The fourth, fifth, and sixth are where the actual personality starts surfacing. By the time you arrive at the version you cannot let go of, the earlier versions are part of why it works.

If your first version looks decent, do not ship it. Push it. The version you ship should be the one that feels like you, not the one that looked good enough on the second pass. I have written about why the Architect was the right archetype once V6 had locked. That post is the resolution. This post is the search.

I am documenting this whole personal brand as I build it, not after it is polished into a case study. The manifesto is the bigger picture if you want it.

Come build with me.

~ Anthony