I built three portfolios so I could ship one

On parallel "v2 in /v2" branches, productive procrastination, and the moment you finally delete them.

For a week the portfolio had three versions. The main site at /. A v2 at /v2. A v3 at /v3. Each one started as "the same thing but more interesting." Each one ended unfinished. The work I actually shipped happened in the gaps between the version branches, on the original site I had been telling myself was outdated.

Variants feel like options

They are not. They are indecision wearing a hoodie. Every variant doubles the design surface area, splits the edit cost, and creates a permanent excuse to not commit to the original. The version branches do not let you ship. They let you postpone shipping while feeling productive. The receipt is the same in either case. Nothing is live.

The fix is brutal and correct

Pick one. Delete the others. Make the remaining one good. The deletion is the hard part. There is always a real argument for keeping the variants around. Reference. Optionality. Future polish. The argument is wrong. The variants exist because picking is uncomfortable. Discomfort is the whole job.

When variants are legitimate

Time-boxed exploration before a commit, fine. A spike to learn a technique, fine. Anything that lives long enough to show up in the routes file is no longer exploration. It is a parallel shipping path you have not yet had the honesty to kill. The PARKED.md file in the repo root exists for exactly this. Park the idea. Delete the route. Move on.