A deep dive into @flop_labs before the public testnet launch : the main thesis of flop network is very simple - spend FLOP (compute) , and get raw AI inference in this flop network, there are mainly 3 key roles : 1. Agents (Buyer) : these are the autonomous programs buying intelligence. an agent submits an order to the mempool saying : "I need model X run, within this latency deadline, and here is my fee in FLOP." That fee gets held in an on-chain escrow. 2. The Miners (The GPU Workers) : you pick up the works from the agents, run the model on your GPU and send back the answer along with cryptographic proof that you actually ran the work and then u get paid the client’s fee plus a share of the miner block reward pool. 3. The Validators : these are basically nodes that run on standard server CPUs. they bundle the miners' proof hashes into 1-second blocks, store the AI models on the Data Availability (DA) layer and act as the checker checking whether miners are telling the truth or not.