Igor

x.com/i/article/2074…

Our next PredictTO event: a fireside chat with Dr. Robin Hanson. Dr. Hanson is the author of the original Futarchy paper, and inspiration behind @Polymarket, leading to the breakout success of modern Prediction Markets. July 29th, Toronto. We're honored to host him 👇

By no means am I an expert in compute markets, but was inspired to develop a take by both Brett and @0xfishylosopher Compute keeps getting scored against the commodity formation checklist and keeps failing the same two boxes, standardized delivery and time plus location pricing. Fair enough as history. Grain, oil and electricity all needed those things because their markets terminated in a dated futures contract, and a dated contract has to bridge into the physical world somehow. But new markets stopped terminating there. Perps collapsed the term structure into a funding rate and swapped the delivery bridge for an oracle print. Hyperliquid became the deepest novel derivatives venue of the cycle on underlyings it never delivers. AX clears over a billion a month in perps on traditional assets and already lists GPU-hour contracts. The catch is that the perp stuffs the whole thing into the benchmark. A funding rate needs an anchor, and for compute the anchor is the entire fight. Compute varies by SKU, location, tenor and counterparty. No index can carry all four, so every index picks what to keep and drops the rest. The three racing to list made three different cuts: on-demand rental rates, printed transactions only, energy normalized. One of these is the LIBOR to SOFR argument, real prints over posted quotes, being made at the birth of a benchmark instead of after a scandal. Whatever an index drops has to go somewhere, and it goes onto dealer balance sheets. Which is why this market is forming dealer first, exchange second. The cleared contract ends up as the hedge leg for the OTC book, and price discovery moves into the spread between indices. Oil did this with WTI-Brent, gas did it with JKM-TTF. When no benchmark is complete, the basis between imperfect ones carries the information. That still leaves the part neither a funding rate nor a basis trade can touch: jumps. An export control ruling, a Rubin slip, an interconnection queue decision, a hyperscaler capex cut. The whole surface moves at once and resolves as an event, not drift. The instrument for that slice is a binary, and the first compute price markets settling on printed transaction data went live in May on @Kalshi. For a dealer warehousing SKU basis that’s the right shape, you lay off the timing and keep the size. Perp for the level, index menu for the shape, dealers for the residual, binaries for the jumps. All four layers are live today. Historical path-dependencies may not be a precise source of truth here.

HYPERLIQUID POLICY CENTER AND TRADE.XYZ MET WITH SEC'S CRYPTO TASK FORCE TO ADDRESS CRYPTO REGULATION: WEBSITE