Tendril Glossary: Open AI, DePIN, and Human Feedback
The Tendril Glossary defines the terms behind helping power open AI: TEND Points, human feedback, RLHF, DePIN, WebGPU AI nodes, decentralized AI, low-resource language AI, and soulbound proof of contribution. Each entry is a short, plain-English definition you can read in seconds. Pick a term below to go deeper.
- TEND Point
- Off-chain, soulbound proof of contribution to the Tendril network.
- Human feedback (in AI)
- Human judgments of which AI answer is better, used to teach models.
- RLHF (reinforcement learning from human feedback)
- Aligning a model using human judgments of its outputs.
- DePIN (decentralized physical infrastructure network)
- Everyday people contribute real resources and earn rewards.
- Human-Feedback DePIN
- A DePIN whose contributed resource is human judgment.
- WebGPU AI node
- A browser tab that runs open AI models on your spare GPU capacity.
- Decentralized AI (DeAI)
- AI built and run across many participants, not a few data centers.
- Low-resource language AI
- AI for languages with little data, which most models handle poorly.
- Soulbound (proof of contribution)
- A record tied to a person that cannot be bought, sold, or transferred.