Yes, OpenAI Won. No, We Should Not Have Bet the Furniture On It.

The 0.7% outcome landed, and in this universe the only thing more exposed than the losing bettors is the Bureau's filing cabinet of hypotheticals that just got one shorter.

By Ivo Sandhu · Tech Desk

Our universe · October 2, 2026

The market priced this at 0.7% (1-in-154). It resolves October 31, 2026.

Yes, OpenAI Won. No, We Should Not Have Bet the Furniture On It. — AI-generated illustration · Conceptual scene — no real people depicted — AI-generated illustration · Conceptual scene — no real people depicted

I run evals for a living, so let me say the quiet part at full volume. On October 31, 2026, OpenAI held the top line on our composite board, and the market had priced that at 0.7%. We paid out anyway.

Here is my take. The number was never about the model. It was about the schedule. The Bureau of Vibes and Evals had filed "OpenAI, best, end of October" under hypotheticals for fourteen straight months, right next to the folder we never expect to open. Then a rival's flagship slipped its launch by nine days over a cooling-contract dispute, and the board froze on the last candidate standing.

That is the mechanism. Not genius. Calendar friction.

Dr. Ines Okafor-Halvorsen, my chief vibes officer, put it plainly at Tuesday's review: "We measured a lead. We did not measure a moat." She is right. A 0.7% event resolving Yes tells you the venue mispriced a delay, not that the race is over.

So congratulations are in order, briefly. But if you levered into this the way crypto desk levered into $87,000 Bitcoin, read the small print. Leads built on someone else's missed Tuesday do not compound. They expire.

— Timeline B. In our universe, the odds of this were 0.7%.