Site Evolution

What I wanted

The intent was not just to make the site dark. I wanted a visual baseline that felt technical, focused, and a little more product-like than a standard personal homepage.

What it changed

That early choice gave the site a foundation I could keep building on: high contrast, quiet motion, and enough surface area for future modules.

What I wanted

I started with a simple browsable list, then kept adding structure: use cases, relationships, filters, and a graph. I wanted the information to feel playable.

What it changed

The lesson was useful. Static lists explain. Interactive systems invite people to look around and form their own map.

What I wanted

The alternative was more one-off page wiring and content logic leaking into places it did not belong. MDX let me keep posts as files while the UI handled layout, metadata, and rendering.

What it changed

I am not an architecture expert, but the product instinct was clear: make the publishing path repeatable before the writing volume exposed every shortcut.

What I wanted

I had planned to come back with a more ambitious animation layer once the site had enough shape. The 3D hero was that moment: a major UI overhaul, not a decorative flourish.

What it changed

It pushed the site away from static portfolio energy and toward something that felt more like an AI-native product surface.

What I wanted

I used a bookmark export to find AI tools I had collected over time and paired that with spending history to separate casual interest from real usage signal.

What it changed

That became the product idea: not another generic tools list, but a tiered view of my own practical fluency and where I had the most reps.

What I wanted

The goal was to create a recurring editorial surface: more opinionated than a directory, more structured than a feed, and better suited to the way I think through AI shifts.

What it changed

The site started to separate modes of thinking: projects, essays, tools, and a more direct point-of-view section.

What I wanted

I added personal photos, family context, and more specific proof because the site should not read like a generic AI portfolio. It should feel like it belongs to one person.

What it changed

That changed the emotional temperature of the site. Less template. More operator, parent, builder, and product person.

What I wanted

I wanted the page to carry the working style: framing problems, owning loops, building thin slices, and using AI as leverage without pretending the tool has the taste.

What it changed

It became the bridge between the personal story and the product thesis behind the rest of the site.

What I wanted

The toggle looked straightforward from the outside. The useful work was finding every place where the UI did not adapt cleanly and tightening the design tokens until both modes held up.

What it changed

That made the design system more honest. A theme toggle exposes shortcuts fast.

What I wanted

LineupSnap uses a fuller backend pattern for invites and registration. This site did not need that level of machinery, so I used Resend for a lighter subscribe and unsubscribe flow.

What it changed

The result was a practical product loop without pretending this frontend-heavy site needed a full application backend.

What I wanted

I wanted to show the sequence of decisions without turning the page into a self-congratulatory changelog. The useful story was what felt wrong, what changed, and what each pass taught me.

What it changed

The release history became evidence for the working style described on the About page, even before I was ready to make it public.

What I wanted

I wanted the portfolio to answer a harder question than “what did you build?” It needed to show why each project existed and what changed in my judgment because I shipped it.

What it changed

Small personal builds moved from decoration to proof. The page started showing a repeatable pattern across coaching, family, product work, and AI-assisted development.

What I wanted

The job was to add observability and newsletter operations in proportion to a personal site: useful when configured, close to zero cost when they were not.

What it changed

The site gained an owned distribution loop, analytics, and error monitoring while preserving the lightweight public experience.

What I wanted

I wanted the whole site to reward specificity. That meant retiring the generic agent catalog, publishing the real AI operating model, tightening the leadership story, finishing two strong drafts, and archiving the one that did not earn a place.

What it changed

The site became smaller in concept and stronger in point of view. The evolution page moved from a review artifact into the public proof of that choice.