Israel

𝗧𝗵𝗲 𝗥𝗼𝗹𝗲 𝗼𝗳 𝗩𝗲𝗿𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗔𝗿𝗼 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 @AroNetwork brings together resources from different locations to support services such as content delivery and AI computing. However, when a network depends on many different nodes, it needs a reliable way to confirm that they are doing the work they are expected to do. 𝘏𝘰𝘸 𝘤𝘢𝘯 𝘢𝘳𝘰 𝘯𝘦𝘵𝘸𝘰𝘳𝘬 𝘷𝘦𝘳𝘪𝘧𝘺 𝘵𝘩𝘢𝘵 𝘵𝘩𝘦𝘴𝘦 𝘯𝘰𝘥𝘦𝘴 𝘢𝘳𝘦 𝘱𝘦𝘳𝘧𝘰𝘳𝘮𝘪𝘯𝘨 𝘵𝘩𝘦𝘪𝘳 𝘵𝘢𝘴𝘬𝘴 𝘢𝘴 𝘦𝘹𝘱𝘦𝘤𝘵𝘦𝘥? To understand how aro network handles this, we need to look at how it verifies the work performed by its network resources. 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 𝘃𝗲𝗿𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗮 𝗱𝗶𝘀𝘁𝗿𝗶𝗯𝘂𝘁𝗲𝗱 𝗻𝗲𝘁𝘄𝗼𝗿𝗸 In a distributed network, resources come from different devices and locations. these resources may have different levels of computing power, internet speed, and connectivity. Bringing these resources together is only one part of the process. the network also needs a way to check the work they perform and establish trust in their contributions. Without reliable verification, it becomes difficult to confirm whether nodes are completing their assigned tasks as expected. 𝗛𝗼𝘄 𝗱𝗼𝗲𝘀 𝗮𝗿𝗼 𝗻𝗲𝘁𝘄𝗼𝗿𝗸 𝗵𝗮𝗻𝗱𝗹𝗲 𝘁𝗵𝗶𝘀? Aro network uses two protocols that play different roles in verifying work and maintaining trust. 𝟭. 𝗚𝗣𝗼𝗪 (𝗚𝘂𝗮𝗿𝗮𝗻𝘁𝗲𝗲 𝗣𝗿𝗼𝗼𝗳 𝗼𝗳 𝗪𝗼𝗿𝗸) GPoW operates at the resource level. It provides verifiable proofs of work performed by network resources, allowing the work to be checked rather than simply trusted. 𝟮. 𝗚𝗣𝗼𝗦 (𝗚𝘂𝗮𝗿𝗮𝗻𝘁𝗲𝗲 𝗣𝗿𝗼𝗼𝗳 𝗼𝗳 𝗦𝘁𝗮𝗸𝗲) GPoS supports on-chain verification, settlement, and governance of work proofs. validators stake tokens as collateral to help secure the network. Together, these protocols help aro verify work performed by its resources and maintain trust across the network. 𝗪𝗵𝘆 𝗶𝘀 𝘁𝗵𝗶𝘀 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁? Aro brings together resources from different locations to support services such as peer-to-peer content delivery and AI computing. For this to work, the network needs more than available hardware and internet bandwidth. it also needs a reliable way to verify the work performed by those resources. By combining verifiable work proofs with on-chain verification, aro network builds a foundation for coordinating distributed resources and delivering services across its network.

AI models learn from examples, so the quality of the information they learn from matters. Some of this information needs to be labelled by humans. this means people review images, text, or other data and add labels to help an AI model understand what it is learning from. 𝘉𝘶𝘵 𝘩𝘰𝘸 𝘤𝘢𝘯 𝘢 𝘤𝘰𝘮𝘱𝘢𝘯𝘺 𝘬𝘯𝘰𝘸 𝘵𝘩𝘢𝘵 𝘵𝘩𝘦 𝘱𝘦𝘰𝘱𝘭𝘦 𝘥𝘰𝘪𝘯𝘨 𝘵𝘩𝘪𝘴 𝘸𝘰𝘳𝘬 𝘩𝘢𝘷𝘦 𝘵𝘩𝘦 𝘳𝘪𝘨𝘩𝘵 𝘲𝘶𝘢𝘭𝘪𝘧𝘪𝘤𝘢𝘵𝘪𝘰𝘯𝘴? This is particularly important in fields such as medicine and law, where mistakes can affect the quality of the data used to train an AI model. For example, medical images may need to be labelled by people with relevant medical knowledge. if someone lacks that knowledge, they could label the images incorrectly, adding errors to the training data. 𝗛𝗼𝘄 𝗰𝗮𝗻 𝗽𝗿𝗶𝗺𝘂𝘀 𝗵𝗲𝗹𝗽? @primus_labs explores how zkTLS-based data proofs can help companies verify whether people labelling AI training data have the required qualifications. This can help companies in three ways: ⇒ Checking qualifications: Companies can verify whether labelers meet specific requirements, such as having a medical or law degree. ⇒ Improving data quality: Assigning specialised tasks to people with the right qualifications can help reduce errors in the training data. ⇒ Building trust: Verifying a labeler’s credentials gives companies more confidence that the people working on their data meet the required standards. 𝗪𝗵𝘆 𝗾𝘂𝗮𝗹𝗶𝗳𝗶𝗲𝗱 𝗹𝗮𝗯𝗲𝗹𝗲𝗿𝘀 𝗺𝗮𝘁𝘁𝗲𝗿 Collecting large amounts of data is not enough to train a reliable AI model. The information attached to that data also needs to be labelled correctly. If people lack the knowledge needed for a task, they may label data incorrectly, giving the AI model examples that do not accurately represent what it is learning. By helping companies check labelers’ qualifications, Primus gives them a way to verify whether people meet the requirements for specialised work. This does not guarantee that every label will be correct. Even qualified people can make mistakes. However, checking qualifications helps companies make informed decisions about who they assign to label their training data. As AI becomes more widely used in specialised fields, making sure the people preparing its training data have the necessary knowledge becomes increasingly important.

Crypto lending often depends on the value of the assets a borrower provides as collateral. but what happens when a borrower’s financial position changes after a loan has been issued? Imagine someone uses ETH as collateral to borrow another asset. at first, their holdings meet the lender’s requirements. later, the price of ETH drops, reducing the value of their collateral. The lender may need updated information to assess the borrower’s position and decide whether the loan’s conditions are still being met. @primus_labs explores how proofs of dynamic assets can support this process. In this approach, borrowers provide evidence of their assets at agreed intervals. lenders can use this information alongside on-chain data to assess lending risks and set loan conditions. The process can involve a few steps: ⇒ Initial assessment: the lender assesses the borrower’s available asset and credit information, alongside relevant on-chain data. ⇒ Regular updates: the borrower provides updated proofs of their assets at agreed intervals. ⇒ Checking for changes: the lending system checks whether the assets have changed within a specified range. ⇒ Taking action: according to the documented scenario, repayment or liquidation can be triggered when the specified conditions are met. Each step helps the lending system assess the borrower’s position as it changes over time. 𝗪𝗵𝘆 𝗱𝗼𝗲𝘀 𝘁𝗵𝗶𝘀 𝗺𝗮𝘁𝘁𝗲𝗿? Many on-chain loans require borrowers to provide collateral. this can limit access to lending for people who have other ways to demonstrate their creditworthiness. Primus explores how combining off-chain information, such as asset and credit details, with on-chain data could help lending applications assess risk and consider more personalised lending conditions. For example, verified credit information could help a lender assess whether a borrower qualifies for different loan terms. The actual terms would depend on the lending application’s rules and risk assessment. Providing proof of assets does not guarantee that a borrower will repay a loan or that their assets will retain their value. The proofs provide information for assessment, while the lending system’s rules determine how that information affects the loan. Primus presents this as a possible direction for DeFi lending, including exploring personalised lending rates, changing collateral requirements, and credit-based lending. The key idea is that lending decisions could take more than the collateral provided at the start into account. Updated asset information and relevant credit data could help applications assess borrowers throughout the lending process.