Decentralized AI: Flock It │ Jihao Sun

An episode of the Thinking On Paper podcast, hosted by Mark Fielding and Jeremy Gilbertson, titled "Decentralized AI: Flock It │ Jihao Sun" was published on April 7, 2025 and runs 48 minutes.

April 7, 2025 ·48m · Thinking On Paper

Summary

OpenAI and Google just upgraded their arsenals, trained on everything from your search history to your creative work. You didn’t opt in. You didn’t get paid. And you definitely don’t own what your data helped build.But what if that changed?Jihao Sun, co-founder of Flock.io, joins Mark and Jeremy to Think on Paper about a radically different model — decentralized AI that trains locally, protects privacy, and rewards users. They get into data ownership, blockchain-based training, and whether federated learning can really compete with centralized giants.If your data is building the next generation of AI, who’s it building it for? And why aren’t you in control?Please enjoy the show. And share with a curious friend. --CHAPTERS(00:08) Introduction to Decentralized AI(01:33) The Importance of Decentralization in AI(04:40) Accessibility and Cost of AI Technology(07:53) Understanding Federated Learning(11:38) Challenges of Centralized AI Systems(14:30) Exploring Federated Learning in Banking(17:10) Blockchain's Role in Decentralized AI(20:45) Democratizing AI Development with Flock(26:49) Real-World Use Cases of Flock(32:56) The Evolution of Parenting and Technology(35:28) Decentralization vs Centralization in AI(36:31) Future of Decentralized AI Models(37:29) Understanding AI Hallucinations(41:23) The Role of Synthetic Data(44:59) Getting Involved with Decentralized AI(45:53) Staying Out of Your Own Way(47:03) Final Thoughts on Decentralization and AI--Learn more:www.flock.iowww.thinkingonpaper.xyz

Episode Description