Same content, same evidence, fifteen different bets about what a portfolio is for. The main site is untouched — these exist to pressure-test it. Each one names its bet; steal what wins.
Fifteen were built; four deserve to exist. The written verdicts were the real output — the other eleven stay up below as the kill list.
The résumé IS the site — with handwritten margin notes redlining it.
The bet: highest hire-signal per pixelVerdict — the real challenger 04The LedgerBrutalist table of engagements: sortable columns, expandable rows with receipts. Giant uppercase masthead, hairlines in full ink.
The bet: density as confidence — a body of work that survives a spreadsheetVerdict — closest to shippable 05The MapA draggable territory. You sit at the center; projects are islands rotated a few degrees off-grid, connected by dashed routes to the regions they belong to. The Compas thesis applied to a career.
The bet: play as differentiation — the portfolio nobody closes in eight secondsVerdict — steal it as a toggle · already live as the site's Map view 01The TerminalThe portfolio as a filesystem: dark, monospace, keyboard-first. A real command prompt, ls-style work listing, images as hover ghosts.
The bet: credibility through austerity — for readers who live in a shellVerdict — easter egg, not front door · press ` on the main site"The Tangled Times." Masthead, columns, drop caps, dispatches. Every case study becomes a headline with a deck.
The bet: editorial authority — work presented as news that happened 03The GalleryNine plates hung in white space, one per viewport, labels fading in as you arrive. The illustrations carry everything; words become wall labels.
The bet: restraint as spectacle — let the image system flex 06The SlidesA 9-slide pitch deck as the site: arrow keys, progress bar, one claim per screen.
The bet: the panel room, skipped ahead to 07The InterviewYou interview Steven: quick-reply chips, typing indicators, pre-written honesty.
The bet: meet buyers in the UI genre they now live in 09The MixtapeProjects as tracks on a cassette deck. Side A: enterprise. Side B: AI-native.
The bet: charm as attention insurance 10The BlueprintWork as engineering drawings — title block, detail numbers, revision notes.
The bet: design presented as rigor 11The PinboardAn evidence wall: draggable exhibits, red string that follows, a person of interest.
The bet: research culture, made visible 12One SentenceThe entire site is one sentence; six phrases open six doors.
The bet: radical confidence in the positioning line 13StevenOSA faux desktop: draggable windows, a dock, and a Trash full of deleted clichés.
The bet: the "personal OS" made literal 14The Wiki"Steven Long Tran — Designpedia": infobox, neutral voice, citations that link to real case studies.
The bet: borrowed genre, instant scannability 15End CreditsThe career as a film crawl — auto-scrolling, hover to direct the speed, everyone credited including the dog.
The bet: the portfolio that ends warmFifteen alternate portfolios — not mockups, working pages with their own interactions — took roughly a day of AI-paired iteration. The 0→1 cost of a direction has collapsed to nearly nothing. That part of the hype is true.
256 combinations means 256 sets of contrast failures, broken hovers, and layouts wearing brands that fight them. Try Poster red on the Wiki, or CRT green on the Broadsheet — the matrix makes bad taste one click reproducible. No team reviews 256 of anything.
The lasting output is the written rationale: the One-Pager is the strongest challenger to the real site, the Ledger could ship tomorrow, the Map should be stolen as a toggle, the Terminal belongs behind an easter egg — and twelve directions were correctly killed on paper.
When generation is free, the craft moves upstream: choosing which directions deserve tokens before you burn them, and stopping when the answer is clear. This page is the argument — and the confession that it took building 256 versions to make it.
../images/, so this folder must sit beside portfolio-site/ (it does, in 12-career). Nothing here is deployed. The taste file was deliberately loosened per direction: 01 breaks the palette (dark, violet accent), 02 breaks the type system (Playfair), 03–05 stay closer to canon. Each direction now carries its OWN brand — palette dots shown on every card; the house palette appears in none of them except where the genre demanded neutrality. Return to the standing site →