License pairwise preference: given the same task, two agent rollouts, and your lived judgment, pick the better one and write the reason. This is the dataset Scale, Surge, Mercor, and frontier labs still cannot synthesize.
This venue is a Wing in your Human architecture. Every exercise creates Rooms organized by Hall type. Data quality is tracked via Knowledge Graph triples.
Meet all milestones to unlock buyer engagement. Hybrid threshold — both content volume AND time consistency required.
Entries
Active Days
Min Days
Agentic companies train reward models on human preference, not on more tokens. A ranked pair with a written rationale is worth more than a thumbs-up. You already do this when you review work.
3–5 hours initial, then 20 min/week · 4 steps · Each step builds your architecture
Preference pairs with rationale (JSON: prompt, A, B, winner, reason, stakes)
When readiness is achieved, an AAAK-compressed summary of your data is generated for buyer evaluation — privacy-preserving, compact, and instantly readable by any AI system.
Largest buyer of human preference and RLHF data for frontier and agent models.
Expert-only preference and critique data for labs that will not accept cheap click-work.
Matches domain experts to AI labs that need preference and eval work.
Constitutional AI and Claude agent products consume human preference and critique at scale.
Open the Data Builder to complete guided exercises. Each entry builds Rooms, Halls, and Knowledge Triples in your Human architecture — reaching buyer readiness in 30–60 days.