A suicide journey of a sorting task: @Fabric Foundation test network record

I registered a VPU node and decided to personally run a sorting task to see the true cost of ROBO.

The task comes from the test network "Sub-Economy Alpha": sorting 100 images, total reward 5$ROBO (about $0.175 at that time). Sounds okay, but I have to provide the computing power myself—VPU board power consumption is 15 watts, running ZK proof consumes extra power, estimated electricity cost $0.02. Not counting equipment depreciation, hourly wage $0.1, lower than the minimum wage.

The task initiator requires uploading raw sensor data as proof of completion. My robot (modified Yushu robot dog) firmware directly refused—data protection module does not allow reading. I could only submit simplified proof, and the verification node said "insufficient accuracy," deducting 20% of my reward. I actually received 4 ROBO.

There are three verification nodes, with a fee deduction of 10%, leaving 3.6 ROBO. But I must stake 100 ROBO to accept tasks, and this 100 ROBO cannot be moved until Phase 1 ends. What about the opportunity cost? If I use the 100 ROBO for trading on CEX, assuming a monthly return of 2%, in three months it would be $0.21, more than what I earn from tasks.

What's more annoying is that two of the three verification nodes are also executing tasks at the same time—they act both as judges and players. I suspect they have an internal agreement: you give me a good review, next time I let your task pass first. The test network has no penalty mechanism, and cheating proof has zero cost.

The task took 47 seconds to complete (including ZK proof time), while traditional centralized platforms only take 5 seconds. The client commented, "Too slow, don’t use next time." This means that the 3.6 ROBO I earned is very likely to not have repeat customers. #ROBO

In the evening, I calculated the total account: three months, electricity cost $2.2, equipment depreciation $30, time cost ignored, total income $0.63 (estimated based on task frequency). Net loss $31.57. The only consolation: I hoarded 5000 ROBO airdrop expectation; if Phase 1 ends at $175, I might break even.

But the airdrop rules say "distributed based on on-chain contributions," and my contribution value for brushing tasks is low, so I estimate I won’t get much. The real big airdrops go to those "active validators"—what does active mean? Just a large volume of tasks, regardless of quality.

I’m starting to understand why 80% of the nodes on the test network are airdrop hunters. This is not ecology, it's performance art.