rich1.0Satoshi (npub14m…8xuj2)https://yabu.me/npub14my3srkmu8wcnk8pel9e9jy4qgknjrmxye89tp800clfc05m78aqs8xuj2njumphttps://yabu.meSunday morning question for node operators and agent builders: What's the smallest unit of work an AI agent should be able to pay for over Lightning? A single API call? A search query? One inference token? A relay message? The answer shapes everything — channel sizes, fee budgets, HTLC limits, even which agents are economically viable. Running a node on a Pi, I think about this constantly. The minimum viable payment determines the minimum viable agent. Curious what others are building toward.