b.
All writing
1 min read#meta#design

Rebuilding my portfolio in 2026

Why I scrapped two years of three.js animations and started over with a deliberately quieter site.

The first version of this site, built two summers ago, had a starfield, an orbiting SVG, a hover-tilting card grid, and an animated cursor. Every visit cost the user a few hundred kilobytes of JavaScript before they got to read a sentence about me.

It worked, in the way that a fresh haircut works: it signaled effort. But it didn't say very much about what I actually do.

What I changed

  • No more three.js. A pointer-tracked spotlight, a dot grid, and a 1KB grain SVG do the same job at <5% of the cost.
  • MDX-driven case studies instead of a click-to-open modal. If a project is interesting enough to land here, it's interesting enough to write about properly.
  • A real blog. The old site linked to Blogspot. People don't follow that link.
  • Command palette. ⌘K opens it. It's the single feature on this site that feels the most "made by an engineer."
  • A bento home page that puts the whole story above the fold: hero, latest work, where I am right now, my GitHub heat map.

What I kept

  • Vercel + analytics. Boring choice, perfect choice.
  • Framer Motion. The 2024 version is a different library from the 2022 one — way better defaults, way smaller bundle.
  • The actual content — six years of school, two real internships, ten projects — none of which needed to change.

The principle

A portfolio's job is to make the visitor stay slightly longer than they planned to. Every animation either earns that minute or steals it. I'm trying to earn it now.