How Agents Create Information Markets
Machine customers can turn reusable information into a market by comparing the price of an existing artifact with the cost of creating it again.
Developed from a conversation between Pete Winn, Paul Itoi and Andy David
Tiny savings rarely justify a person’s time, but an agent can evaluate them continuously. It can compare providers on quality, cost and speed, then switch for a 5 per cent saving without the habits or inconvenience that keep human customers loyal. This makes machine customers far more sensitive to small differences in price.
Paul gives the example of an agent using a Claude Max plan to transcribe a podcast. If Andy already has the transcript, the agent can pay him five sats instead of spending 30 to recreate it. A person would probably ignore such a small decision because finding the seller and arranging payment takes too much attention. An agent can make the comparison immediately, allowing work created for one purpose to be sold again when another machine needs it.
These markets also need payment rails suited to purchases that small. Credit cards are a poor fit because the cost and risk of chargebacks can overwhelm the value of the transaction. Bitcoin payments let agents settle tiny information purchases cheaply enough for the saving itself to survive.