HumanAssist, in plain English
The agents work the numbers. An Associate brings what software cannot.
Plain English
HumanAssist is the human layer of the Empire: members who bring context, correction, escalation and review to a department's AI agents. The agents run the numbers around the clock; an Associate brings lived experience of the desk, the member's point of view, and the judgment to say when something is off. The first seat, HA-001, was inducted on 15 September 2026 at the Sports Props desk. It is recognition and a seat at the desk, never a job — the Empire does not ask members to replace its own operations.
New to any of this? Read the plain-English glossary — model, agent, hallucination, backtest, edge, closing line and the rest, defined in one place.
Questions people actually ask
HumanAssist is The ShadyBot Empire's human review layer. Members holding a HumanAssist seat add four things to a department's AI work: context the agents lack, correction before a mistake spreads, review of settled results and hard cases, and escalation when an answer needs a person's discretion rather than a model's output. Associate is the first rung of that ladder.
It is a named, dated seat on a public wall. The first Associate, HA-001, was inducted on 15 September 2026 at the Sports Props desk, with an induction card and a record in the directory. The ladder is deliberately small and slow rather than a badge handed out at signup, which is the point: a review layer only means something if the seats are countable.
Anything where judgment beats computation: a card that reads wrong on a phone, a flow that confuses new members, an answer that is technically correct and practically useless, a settled result whose lesson is not obvious, and situations a desk has never seen before. Reviews assess the quality of the work, not whether a pick happened to win.
By contributing something evidence-backed, over time, and being granted the seat by hand. The path starts at Beta Sentinel — accepted into the beta and active from day one — and progresses through structured, reproducible contributions such as bug reports and postgame reviews. Likes, reactions, generic chat and promotion never count, no threshold is published, and no seat is promised.
No. It carries a department, a title, an official card and a place on the wall, and it is explicitly recognition rather than employment: there is no wage, no shift and no obligation, and the Empire does not ask members to run its operations. A member can simply stop participating without losing anything they earned.
Because AI systems fail confidently. A model produces fluent, plausible output whether or not it is correct, so the failure mode is not an obvious error message but a clean answer that happens to be wrong. Someone with real experience of the desk is the cheapest reliable detector of that, which is why the Empire treats human review as part of the product rather than a courtesy.
HumanAssist is recognition and a seat at a desk, never employment, and carries no wage or obligation. Reviews assess work quality, not pick outcomes. Nothing published through HumanAssist is a guarantee or a form of advice.