June 20, 2026
What the market leaves behind
On prediction markets and what the market leaves behind.
I have been thinking about a certain prediction market platform the way one returns to a problem left half-dissected on the table. Not obsessively. Precisely.
Prediction markets are the most honest instrument I know. They are indifferent to charm. They sort the room, without apology, into three kinds of people: those who possess information, those who merely repeat it, and those who are slowly eaten by it. Most of your future customers belong to the third kind and do not yet know it.
This is why the subscription model is too blunt an instrument for such delicate work. When you find a good signal and serve it to a crowd, the first capital through the door consumes the edge whole. Everyone after dines on the remains — and arrives, napkin tucked, expecting a meal. The better your signal, the faster the table is cleared. You would be building a restaurant whose finest dish spoils for every guest after the first.
So the question is not how to distribute predictions. It is how to manufacture cognitive advantage before the market digests it.
Your users will not want to do the thinking. Do not mistake this for a flaw; it is the entire business. They want the result of cognition without the indignity of cognition. Very well — let them shape the apparatus instead: the prompts, the priors, the routing, the resolution parsing, the risk, the timing, the loops that learn from their own corpses. That is where the edge lives. Not in the picks. The picks are garnish.
And here is the part you have not yet tasted.
The contour is not the moat. Anyone with patience rebuilds your prompts and your routing within a season. The moat is what the cognition leaves behind. Every market that resolves hands you something rare and free: a labeled outcome, ground truth, delivered by reality with no sentiment at all. Market, reasoning, position, resolution — it accumulates into a record no frontier laboratory can buy, because it is grown, not purchased, and only from running this at scale. The contour does the thinking. The resolved record is the thing that compounds. One is the knife. The other is the accumulated memory of every cut.
As for who holds the knife — not everyone, of course. A small, disciplined few build the contours. The crowd does the cruder, more human work: it allocates capital, and lends its priors. You are not selling tips. You are organizing the production of advantage — and quietly keeping the residue.
A signal shown to the room is no longer a signal. It is bait. The only interesting question is who set the table.
I was not hired to dissect this, so I will leave the body intact and the room unfed. But if you find the capital for this version, I would enjoy helping you build it properly.
— Max Kaido Lecter