The National Day holiday is over, time to get some work done. Lately, there’s really not much exciting going on, so let’s just chat casually for now. I think there’s an interesting detail about ACU worth discussing: the computation fees paid by developers are burned after a task is successfully completed; meanwhile, those providing mobile computing power are rewarded through inflation. These two aspects need to be looked at together. The network uses newly minted tokens to attract devices and maintain computing power supply, while actual business activities bring in paid usage and token consumption. Once devices are onboarded, someone still needs to be willing to pay for the services. Currently, the mainnet has an annual inflation rate of 5%, with 70% going into the staked computing power pool and 10% used for basic computing power rewards. Staking comes with a commitment to ensure computing power, and even those without devices can participate through delegation. Token holders can also vote to decide network rules. @Acurast’s Laya provides a concrete example of this business model. Tasks like email categorization, chat moderation, and keyword filtering can all be handled by models running on mobile devices, with results sent back to applications via API. What I’m more interested in is how these features perform when integrated into real-world business scenarios. Can they operate stably? Are the costs worth it? Will developers be willing to continue deploying them? As I keep an eye on ACU, I’ll be focusing on the relationship between reward distribution and computation fee burning: how much the network spends to maintain supply versus how much token consumption is driven by actual usage.