Harry phan
@harrypy
Harry phan
@harrypy
lock in
lock in
Hồ Chí Minh, Vietnam
Devfolio stats
Devfolio stats
2
projects
2
2
prizes
2
3
hackathons
3
0
Hackathons org.
0
Top Projects
Top Projects
With AI's growing presence across sectors, the promise of addressing complex problems and streamlining operations holds immense appeal. However, a significant hurdle looms in the background - AI 'hallucinations'. These inaccuracies arise when AI models make misguided interpretations or predictions based on unverified assumptions, much like perceiving phantoms where none exist. And that's why Wedata was created. Wedata offer a solution to store data on- chain. Pairing this with end-to-end encryption ensures data integrity when shared via the InterPlanetary File System (IPFS) for consumers. Wedata aims to revolutionize the AI industry through blockchain, fostering trustworthy AI and reducing AI hallucinations.
We propose to create a browser extension that uses AI to determine the semantic structure of the websites the user visits on the fly, parse them into semantic chunks on the decentralized storage.
The context We all know a key concern that AI agent learning from unreliable information. So data quality is crucial since data serves as the lifeblood for AI agents. > the more high-quality, verifiable data you have, the smarter and more reliable the AI models can become. The problems But data in web2 is messy and unorganized, with mixed content types. Website owners also don't want to share their data with scrapers. That’s problem make us want to solve this. Our solution This is where AICrawler comes in. It's a tool that helps our community collect and organize web content. As it gathers information, it creates organized patterns (called Semantic Schemes) that make the web easier to understand - both for people and machines. We need to combine AI crawlers with a good reward system. This would create a better way for everyone to work together in collecting and organizing web data. With Move agent kits and FlyFish AI Agents, AI only learns from data that's been verified as trustworthy.This means you're not just making AI smarter - you're helping create a future where we can all trust AI more.