GM ai

Sandeep is one of the smartest people I've gotten to know (also super humble)

We asked the Sandeep Chinchali about the role of IP in the age of AI. "Essentially all public data sets have a non-commercial license. "Sometimes generative AI models that create imagery could have been trained on IP unsafe data and they cannot even use synthetic data." "What

What flows like water 🔱

AI is moving beyond the browser and into the real world. The bottleneck? Data. Today we’re announcing a $15M seed round led by a16z crypto to build infra that collects, curates, and licenses high-quality data for physical AI. Incubated by and built on Story.

<Poseidon: AI’s Full-Stack Infrastructure for IP-Cleared Training Data> Today, Poseidon is launching: a new infrastructure layer for AI, incubated by Story and backed by $15M from a16z crypto, built to fix its most broken input: training data. Nvidia is valued at over $4

I’m genuinely surprised to see a16z quadruple down on a single project, but it makes sense if they view Poseidon as the catalyst for $IP. Poseidon sources and curates high quality datasets, tokenizes them on Story for full provenance, and connects suppliers with AI

Imagine never folding laundry again. The internet trained LLMs, but robots and self-driving cars need training data that is much more challenging to source. Poseidon enables the collection, curation, and licensing of real-world data to accelerate physical AI. Use cases:

🔱

it's crunch time

While building robotics models at Stanford, Sandeep Chinchali saw a pattern: It wasn’t the 99% of data that dramatically improved performance. It was the rare 1% – edge cases and long-tail data. Poseidon is built to make that data usable for AI teams. More on the blog:

Build platforms that optimizes for P2P over PVP

Sup

Sup LJW.ip

Enjoy the season you are in

Eleven Labs crushing it

Keep going