The Collective Intelligence Protocol for AI thats uncensored, unlocked and user-owned - like it should be
Open AI Infrastructure for user-owned AI Pipelines
After years of pushing the boundaries of decentralised AI, we're sad to announce that Lilypad is closing.
We set out to prove that open, verifiable compute could change how AI is built - and we firmly showed it’s not just possible - but advantageous.
Latency Isn’t Just a Tech Stat
⚡ Latency = how fast a model responds.
It’s not just for devs. It affects:
• Real-time robotics
• Live AI agents
• Inference in the browser
• User satisfaction in chat
Even if two models score the same in accuracy…
The one that responds
LLMs make headlines.
Specialized models make progress.
In fields like biotech, logistics, and law, models aren't flexing general knowledge, they’re lifting specific weights:
• Is this mutation significant?
• Is this satellite image showing flood risk?
• Is this transaction
From script to moving image without a camera.
Text-to-Video models generate video clips from written prompts. They're early-stage but rapidly improving and already unlocking real value for creators.
Whether you're prototyping an idea, pitching a concept, or experimenting with
