🤖 2D Agent Plan-Act-Critique Retry Loop: Absorbing Markov Chain, Attempt Budget, Silent Failures and Token Cost
Side view of an LLM agent loop (Plan, Act, Critique, retry): how critic recall and false-accept rate, the attempt budget and per-stage time and tokens set the chance of a correct, silently wrong or abandoned result, verified by a geometric series, a Markov-chain solve and Monte Carlo.
⚙ Under the hood
Side view of an LLM agent loop (Plan, Act, Critique, retry): how critic recall and false-accept rate, the attempt budget and per-stage time and tokens set the chance of a correct, silently wrong or abandoned result, verified by a geometric series, a Markov-chain solve and Monte Carlo.
LLM AgentsPlan-Act-CritiqueRetry BudgetAbsorbing Markov ChainCritic Error RatesLatency and CostAgent ReliabilityMonte Carlo
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install