The weekly BuildersHub has wrapped up again. For those brothers who missed the live session, here’s a quick rundown of the two core projects this time—both are practical and implementable, not just empty promises or pie-in-the-sky ideas.
The first is decentralized storage + AI knowledge base. You can think of it as a more Web3-flavored Notion + private AI.
The logic is pretty straightforward: you upload your texts, images, videos, etc., and the data is stored in a decentralized manner, not on a company's server. This is quite attractive to many people, at least you don’t have to worry about the platform malfunctioning and restricting your content or using it to train other models.
And here’s the key point: its AI is based on your own data. Honestly, many are working on private corpora + AI nowadays, but if they can refine the experience well, it’s useful. After all, general models like ChatGPT sometimes spout nonsense seriously; feeding it your own data makes the relevance much higher.
They also discussed some technical aspects, such as integrating Rialo later for privacy layers and Shelby protocol for encryption and timestamps. It sounds like they’re heading toward making data usable but not visible—pretty mainstream narrative these days.
Personally, I think whether it’s worth using depends not on the words “decentralized,” but on two points: first, whether uploading and managing data is smooth; second, whether the AI’s actual performance is noticeably better than existing tools. If it’s just a shell change, most people won’t migrate; ultimately, these products compete on user experience.
The second project is infrastructure for insurance clearing, which is clearly more industry-focused. Ordinary users probably won’t use it, but it’s solid and real.
Traditional insurance clearing processes, to put it bluntly, are Excel + email hell, with multi-party reconciliation dragging on so long you start doubting your life. Their solution essentially moves the entire process onto the blockchain, using smart contracts for automatic execution.
Here’s a compressed version of the process: the underwriter first uploads the product → other participants join proportionally → each locks funds on-chain → user purchases insurance → an event occurs → oracle confirms → automatic payout. The key is the “automatic clearing” part—reducing human intervention and disputes.
They’re planning a major shift from Solana to Rialo, cutting out some off-chain relay logic, aiming to run the entire clearing process fully on-chain consensus. Plus, with encrypted computation, they can handle sensitive data, which is crucial for heavily regulated industries like insurance.
Another point I think they’re doing well is managing oracle risk. Instead of relying on a single data source, they use multi-source verification. If data conflicts, they pause the process immediately—no further execution. This design is slower but reduces the risk of major accidents.
Of course, a natural challenge for this kind of B2B infrastructure is that, no matter how good the product is, the industry has to be willing to adopt it and willing to change existing processes. The difficulty isn’t technical but business and trust-related.
Overall, these two projects seem quite promising: one focuses on personal data + AI, which, if refined, could break out of the niche; the other is a foundational industry infrastructure—slower to develop, but with higher barriers once operational.
From my perspective, current Web3 projects that still talk about grand narratives but lack concrete implementation paths can mostly be skipped. Those with visible use cases, even if early, are worth observing a few more rounds. No hype, just watching how they iterate in the future.
#CNUGC @RialoHQ
