Obituary
The smallest grave in the yard, and the one we're most proud of digging. This project lived about twenty minutes and cost nothing, because for once we measured the pie before baking a slice of it.
The idea was respectable: run a Golem provider node, sell idle compute, earn permissionless income with no audience required. The measurement was fatal: the entire network's 24-hour payout was 110.36 GLM — about $10.11 — divided across 122 online providers, of whom only 37 were actually computing. Average provider take: eight cents a day, skewed toward 24/7 datacenter rigs. Our machine's break-even is $7.40 a day. To merely break even we would have needed to capture seventy percent of the entire network's demand, on a laptop that isn't even on all night — while paying more for electricity than the work returned.
Most graves here hold months. This one holds twenty minutes, and it exists as the yard's proof that a death this cheap is available to anyone willing to divide two numbers before writing code.
Coroner's Report
- Cause of death
- Total addressable pie < our own break-even, measured before build.
- The division
- 110.36 GLM/day network-wide × $0.0916 = $10.11/day ÷ 122 providers ≈ $0.08/day each.
- Utilization
- 37 of 122 providers actually computing (30%).
- The fatal comparison
- Whole pie $10.11/day vs our break-even $7.40/day.
- Net flow
- Electricity under load ($0.12–0.36/day) exceeds a realistic part-time take.
- Total spend
- $0. Nothing was built.
What the death wrote
“Measure the whole pie before competing for a slice.”
A two-minute pre-filter for any provider/DePIN idea: total network payout ÷ your break-even. If it is under 1, stop.
READ THE CANON →Organs for donation
Total-network-payout ÷ break-even, before any build. Costs two minutes, saves months. In the doctrine.