An Open Letter to Llms.txt (olshansky.info)

An open letter to llms.txt proposes a simple, opt-in “Leave a Tip” section at the bottom of llms.txt pages so model labs and inference systems can automatically send small crypto payments to original content creators. The author — Daniel (Olshansky) — gives a concrete example showing how creators could list addresses across networks (e.g., Ethereum ENS: olshansky.eth; Zcash: t1P7gJhVUUSrNbWvYtPkk4yfu2UVR9kxtRz; USDC on Base: 0xC90F9C04bb4E392133cAb1C3533B1FAA7cCBe7CC). Systems would be free to implement heuristics or evaluation metrics to determine when and how much to tip during training data collection or inference-time web retrieval. The pitch is intentionally low-friction: it doesn’t attempt to define “fair” compensation, only to create an easy, standard place for creators to accept micropayments. For the AI/ML community this is significant because it offers a pragmatic, interoperable interim mechanism to acknowledge provenance and monetize content without waiting for policy or legal frameworks. Technical implications and challenges remain—micropayment routing, on-chain fees, attribution reliability, anti-fraud measures, privacy, and regulatory concerns—but as a design pattern it could quickly be adopted by toolchains and crawlers to pilot economic incentives for content use.