AI is taking over one day at a time but at the root of it, we need more valuable data. AI models need more real-world data but it feels like most has been dried out. It's tough finding all needed in publicly available databases. That's what Poseidon will be focused on. (1) easy way for anyone to provide data (2) proof of ownership (3) an opportunity to earn The flow is simple: > upload data > it gets checked and verified > AI companies can license it > contributors get paid I love the idea behind it and I agree with it. In theory, collecting more valuable data should lead to better results. I’m excited to see this put into practice! It feels like we are one step closer to more successful AI development 🫡
S.Y. Lee Story/IP
S.Y. Lee Story/IPDec 8, 22:27
AI can't go global without global data. To train physical-world models, you need data from Seoul, Jakarta, São Paulo, Nairobi, and more, not just Silicon Valley. Crypto makes that possible. For instance, Story registered ~300K data shard IPs daily from our ecosystem project @psdnai's data crowdsourcing campaign. In ~2 weeks, we had 405k contributors and 34k hours of rights-cleared audio to train new AI models on. This is a preview of the next era: AI trained on real environments, real voices, and real global participation. Story & Poseidon will shape AI’s data trajectory the way NVIDIA shaped the compute layer.
I'm happy to have partnered with Story Protocol for this post 🙏🏼
1.84K