SY Lee

Trace reaches >25M registrations, helping AI labs buy data with confidence. Billions more data registrations to come!

just crossed 15M receipts on Trace! on track to 100x this within a year.

Last week, we became The DATA Foundation. Here's a TLDR if you missed it: > a year ago we set out to build the IP layer for the internet. We made some good progress: > Magma's 3M artists registered and licensed their art on Story. Aria brought music rights onchain, including a $100M Korean music IP tokenization pipeline and rights from Justin Bieber, BTS, Miley Cyrus, Dua Lipa, and BLACKPINK. > but we also learned important lessons along the way, including a hard one we'll own clearly: companies behind the most valuable music, games, and brands guard their most valuable IPs jealously, and the open nature of Story's permissionless licensing clashed with the very control those companies want to keep. > despite this, one direction became the clear frontrunner: AI training data. In 2025 we incubated an AI data project called Poseidon. @psdnai proceeded to raise a $15M seed from a16z and showed immediate traction with frontier labs. > every major lab like OpenAI, Anthropic, Google needs answers to the same three questions: can you source the data we need at scale, can you prove where it came from, and can you guarantee the quality? No one can answer all three. It's a multi trillion $ gap. So we're going all in and dedicating everything to it. DATA’s ecosystem is addressing all three: sourcing data through ecosystem apps like @useKled, processing data through Poseidon, and being the audit trail for human contributed real world data through our flagship product, Trace. i.e. every record will have a verifiable receipt. Kled has over 1.5B human contributed data records that are being registered to Trace over the coming months, where they'll be fully auditable - something they need secure additional contracts with labs. Excited to continue to shape the Trace product with @avipat_ and the rest of the team. LFG

More big updates. Korea's biggest fintech giant, @toss__official, is bringing its 30M users into the AI data economy with @psdnai in its first ever AI/Web3 partnership. Numo will soon be a mini-app within Toss. markets.businessinsider.com/news/currencie…

Big News. Korea's biggest fintech giant @toss__official is bringing its 30M users into the AI data economy with @psdnai, in its first ever AI/web3 partnership. We'll begin working to integrate our contributor app Numo as a mini-app inside Toss, so anyone can provide real-world data to train AI and get paid for it ↓

New name, more focused mission, tackling the most important problem in AI today. We started Story to move the world's IP onchain. What became crystal clear is that the most valuable IP in the world of AI is actually data. Specifically the real, human, licensable data every AI lab needs and can't get cleanly. Soon enough, every company sitting on human behavior will become a data business, whether they intend to or not. The teams that can source that data at scale and prove where it came from with consent will build some of the most valuable companies of this era. Excited to have @devrelius lead as the Chief Executive Officer with @avipat_ as the Chief Data Officer role (advisory) at The DATA Foundation. I'm excited to continue supporting The DATA Foundation team and their mission through my efforts at @piplabsxyz and @psdnai. We won’t stop until all AI training data leaves a trace. We won’t stop until all AI training data leaves a trace.

The bottleneck in AI moved from architecture to data. Voice is the clearest proof. The new speech-to-speech models can finally hold a real conversation but they need a kind of data that barely exists in the wild. Today we're partnering with @otodotearth and @StoryProtocol to fix that. ↓ x.com/psdnai/status/…

@joshelman @a16z Congrats!

1/ I’m thrilled to announce that I’m joining @a16z as a partner to help find and support building the next generation of great consumer companies. Let me talk a little about why.

@StoryProtocol So excited to be a part of building this! Unfortunately I can't participate, I don't know those languages (and i'm not trying it live in office hours again...) 😂 If you haven't already, complete your tasks at numolabs.ai :)

After incubating @psdnai, our conviction only got stronger: The data AI needs isn’t on the internet. Critically, we learned what it takes to collect it: real participation, clean processing, and systems for consent and provenance. Numo is what’s next. Welcome to early access.

It was a huge honour to meet with President @Jaemyung_Lee in Seoul. Deeply appreciate and impressed by our thoughtful exchange about AI safety and the importance of using AI to advance science. Korea has a leading part to play in that, and we look forward to working together!

Excited to speak in front of @POTUS @realDonaldTrump alongside other fantastic speakers.

Cost of generating a Hollywood-grade movie: collapsing toward zero. Cost of figuring out how to pay for the synthetic @neiltyson: undefined. The art is here (!!), the bottleneck is now IP remixing infrastructure—and it's worth roughly all the money.

Every major shift in crypto bridges back to the real world. @Ray_scale from @Grayscale on why Story matters ↓

the most valuable data is the data no one wants to openly share. training data for robotics, proprietary datasets, financial records, medical data, model weights, i.e. the stuff that actually matters will always be mediated through confidentiality and legal agreements. to build smarter models in a decentralized way, we need to build the pipes that let confidential data flow between parties who don't fully trust each other, or don't trust each other at all. we need to do what protocols like zcash do for token transfers, but for actual proprietary data, with licensing and confidentiality built in. this is the problem we've been building toward at story. we started with open IP infrastructure... the idea that digital assets need programmable rules attached to them and anyone can remix them permissionlessly. But the deeper we went, the more we realized there was another massive underaddressed problem: enabling the secure transfer and computation of data that can't be public. That is IP too, and the most precious. not everything needs to be transparent. some things need to be provably private and conditionally accessible. the blockchain industry has spent a decade optimizing for openness. the next decade belongs to whoever builds the best infrastructure for confidentiality. More than just private transactions. as Confidential Data Rails (CDR) heads towards testnet soon, it will change what story can do at its core, supporting high stakes, confidential data and IP natively, which is the precious fuel AI runs on.

@jyseo_cv Love this!

What if a world model could render not an imagined place, but the actual city? We introduce Seoul World Model, the first world simulation model grounded in a real-world metropolis. TL;DR: We made a world model RAG over millions of street-views. proj: seoul-world-model.github.io

Jensen: "OpenClaw is probably the single most important release of software probably ever"

Great to see Story on there. we've been pioneering how agents interact since 2024 starting with our research on Agent TCP/IP because we knew it was the future. We were also the first to demonstrate two agents engaging in a complex transaction, ie licensing each other's training data using Story's programmable IP license. We were the first team to hire an agent, Luna, to run our twitter account for a week (paid gig!), and we were also one of the first teams to have a section of our docs entirely dedicated to agents to consume and learn how to build on Story via MCP. In other words, we have been agent-native for more than a year at this point. We also launched skills and a whole LOTS more cooking!