AI training data is already a multi-billion dollar market, with companies like @scale_AI , @HelloSurgeAI , @mercor_ai and others generating massive revenue by supplying data and environments to AI labs. We’re building Digg, a workflow data network for the next phase of AI training. Our first product, Digg Code, works alongside Claude Code, Cursor and other coding agents/harnesses to capture user-approved developer workflows: task context, tool calls, code edits, tests, failures, corrections and outcomes, then structure them into training data for coding agents. We’re also bringing consent, provenance, data rewards and licensing records on-chain on @solana , so contributors can have a verifiable record of what they approved and participate in the value their data creates. Real workflows → training data → better models and agents. Planning our community raise on @futarddotio soon. More on Digg Code this week.