Most Certainly Try Labs Public experiments, useful weirdness

White Paper + Party Lab

Can AI make Knicks fans and Spurs fans friends?

A fast cultural prototype for turning playoff heat into better questions, sharper jokes, and a little less sidewalk nonsense.

Try the fan friendship scout Read the white paper

Executive Summary

Yes, probably. But not by making anyone neutral.

AI should not sand the edges off fandom. Knicks fans should still sound like Knicks fans. Spurs fans should still sound like Spurs fans. The useful question is whether AI can help people keep the loyalty and lose the worst behavior.

This lab proposes a small, consent-based fan friendship experience. It does not decide who is right. It translates rivalry into shared values: patience, pain, role-player love, beautiful basketball, civic pride, and the complicated pleasure of caring too much.

The Problem

The playoffs are giving us a real-time case study.

Recent coverage has a clear pattern: huge Knicks energy around Madison Square Garden, an ascendant Spurs team built around Victor Wembanyama, and a Finals atmosphere hot enough to turn fan identity into a public safety issue. The cultural opportunity is sitting right there.

Sports culture usually rewards the fastest insult. AI can help create the better second question. Not, "Why is your team trash?" but, "What kind of basketball pain made you this way?"

The Method

Use AI as a translator, not a referee.

The fan friendship model maps five kinds of signal:

  • Basketball aesthetic: defense, pace, passing, shot-making, chaos, systems, and player development.
  • Emotional style: loud, skeptical, patient, fatalistic, analytical, superstitious, or loyal against evidence.
  • Historical memory: dynasty years, rebuilding years, the 1999 shadow, and the feeling of being back.
  • Player archetype: stars, glue guys, defensive sickos, irrational bench legends, and coaches' favorites.
  • Rivalry compatibility: whether the person can talk trash and still be a decent hang.

Recent Playoff Context

Five pieces this lab is reacting to.

MySA: NYC Mayor responds after Spurs fans harassed

Useful because it frames the fan-conduct problem directly: passion is fine; intimidation and violence are not.

NBA.com: Knicks, Spurs add a new chapter to MSG lore

Useful because it captures why this matchup feels bigger than a series: arena mythology, city identity, and the return of Finals stakes.

NBA.com: Game 3 was the most-watched Finals Game 3 since 1998

Useful because it proves this is a broad cultural moment, not just a niche basketball argument.

NY1: Knicks fans celebrate outside MSG despite watch-party crackdown

Useful because it shows the community side of fandom and the safety concerns that come with scale.

The Guardian: Wembanyama's Spurs knock Timberwolves out

Useful because it shows the Spurs story before the Finals: Wembanyama, growth, and a team becoming dangerous ahead of schedule.

Design Principles

How to build the experience without making it corny.

  • Keep the rivalry alive. Friendship is less interesting if everyone has to pretend not to care.
  • Make the prompts specific. "Who is your favorite player?" is fine. "Which role player made you irrational?" is better.
  • Use humor as the bridge. Fans trust a good joke before they trust a lecture.
  • Do not collect more data than you need. The point is connection, not surveillance with confetti.
  • End with something people can say out loud. The output should start an actual conversation.

Conclusion

AI can make the first better question easier.

Can AI make Knicks fans and Spurs fans friends? It can help. Not by declaring peace. Not by awarding moral victory. Not by flattening either fan base into bland sportsmanship copy.

The better use is smaller and more human: find the shared signal, name the tension, give people a smarter opening line, and let the rivalry become a conversation instead of a shove.

Build Notes

How V1 of the app was built and published.

This version is intentionally static: one HTML file with embedded CSS and JavaScript. That keeps the party demo fast, cheap, resilient, and privacy-light. No account system, no database, no silent collection of fan feelings.

  • Research tools: Web search and source review were used to pull recent playoff context from MySA, NBA.com, NY1, and The Guardian. Those links became the white paper's reference spine.
  • Writing tools: Codex helped turn the research into the white paper argument, keeping the tone specific, practical, and a little mischievous instead of generic sportsmanship mush.
  • Build tools: HTML, CSS, and vanilla JavaScript were used for the live app. The scout calculates a playful compatibility score, writes a short truce readout, and generates conversation prompts entirely in the browser.
  • Validation tools: Python's standard HTML parser checked that the new page and Labs homepage still parse. `rg` verified the expected sections, links, and app text were present before publishing.
  • Publishing tools: `git` committed the new lab page and homepage link, then pushed to GitHub. GitHub Pages publishes the `main` branch to the public Labs site.
  • Cloudflare/custom domain role: `labs.mostcertainlytry.com` is the public custom domain. The repo's `CNAME` file points the Labs site to that hostname, and DNS resolves the subdomain to GitHub Pages. V1 does not need Cloudflare Workers, D1, KV, or R2 because nothing is stored server-side.
  • Cache check: After publishing, the live page was verified with a cache-busted URL so the new version could be confirmed even if an edge/browser cache was slow to refresh.

V2 option: If the app needs real submissions later, the clean next step would be a Cloudflare Worker endpoint with D1 or KV storage, explicit consent, spam protection, and a simple admin/export view. For tonight, browser-only is the safer demo.