Frank and I built First Press because we believed that how people engage with music could be turned into something with real economic value. Not streaming. Not social. A game where fans wagered on their taste and knowledge, and artists shared in the rake.
We started by building daily, hosted prediction games. Players listened, placed small bets, and tried to predict the crowd’s favorite track, or the host’s favorite. At their best the games were Wordle-like: shared, daily, hard enough to feel rewarding, easy enough to come back to. In an early game with fewer than 50 players and no wager over $11, the winning artist took home more money in a single day than they would have made in a year through streaming.
We knew the road to legally operating a money in/out game was long and expensive. When Kalshi had secured legal status and Polymarket was close behind, we saw a faster route. We pivoted from building skill-based money in/out games to building the data layer for platforms that already had the license and the users.
Our insight was that unlike sports, culture had no score, and no scoreboard that we watched collectively. Culture resolves on award cycles and year-end charts, slowly, infrequently, controlled by insiders. We believed that if we built trusted, tradeable scores for cultural artifacts — books, films, songs — we could create our own resolution events. While other prediction markets waited for outcomes, we would create them.
Then we met with the platforms and learned two things:
First, the architecture of prediction markets is built for mass-scale, synchronized resolution. That’s not a design choice. It’s a structural constraint. Culture may have worked in the pre-streaming era, when radio and MTV meant more frequent, mainstream shared experiences. But streaming and then personalization fragmented all of that. Our belief was that thousands of niche markets could replace one water cooler; the aggregation of niche could be the new reach. What we learned was that the operational overhead is the same for a market with 100 traders as 10,000. You can't solve that by running thousands of smaller markets. The math doesn't work. Certainly not now. Maybe not ever.
Second, the best markets run on fervor. This was a word one of the platform operators used that haunted me for weeks. In the best markets, he said, you don’t just want your side to win, you want the other side to lose. Sports fans understand this intuitively. Even my 9 year old knows that to love the Seattle Seahawks means hating the San Francisco 49ers. That opposition is structural, not incidental.
Culture generates passion, not fervor. The difference matters. Passion pulls us toward something, and each other. Songs, films, fashion, dance are social objects, divinely designed to bring us together, even when we disagree about what we love or hate, what’s merely good vs. great. Prediction markets aren't just built around fervor. They manufacture it. "Bet on anything" is the promise, but should we? We weren't willing to architect opposition into culture, no matter how lucrative the rake.
Here is the part I keep coming back to.
Is there a community-specific version of this that works precisely because it’s small and synchronized? Certain fan communities and friend groups already have passion, shared attention, and even adversarial conviction. They travel together, argue together, and want to be right about which version of which song their favorite band will play that night. The infrastructure doesn’t exist for fan communities to build their own prediction games, or markets, but what if it did? (Ps. PhishMarket.com is only $1,000)
And this: could trusted scores for cultural artifacts (what we called Truth Files) unlock value for fans and creators outside the prediction market context entirely? The score as artifact, not as instrument of a bet, may be the more durable idea, especially as distrust in algorithmic feeds grows and content provenance, like the nutrition labels on the food we eat, becomes something we care more about in the imminent era of synthetic media.
For now, we’re putting First Press on the shelf. We’re proud of what we built and learned. We’re sharing this because the questions we were chasing still matter — even if we’re not the ones who’ll answer them. We’re still asking. We’d love to know if you are too.