1\section{Conclusion}2\label{sec:conclusion}34We set out to identify, inside a transparent exchange economy, the causal mechanisms through which5autonomous language-model agents change price discovery as their share of the market grows. The6instrument is an event-sourced limit-order-book simulator with exact accounting, a latent dividend7process that keeps fundamental value, public value and price apart, calibrated classical traders, and a8randomized, matched-seed design in which the LLM share is the treatment. The theory predicted that a9shared model would degrade informativeness through correlated errors, raise coordination, amplify10leverage spirals, and that diversity would help. The experiments confirmed the coordination and diversity11predictions and reversed the price predictions.1213Three findings stand out. First, in this economy a small reasoning model trading from a canonical public14snapshot is, overwhelmingly, a patient liquidity provider anchored on the public present value: spreads,15impact and volatility fall and depth rises with its share, with no bubble or crash in any live run, and16with leveraged traders present the LLM agents absorb the fire sales that destabilize the classical market.17Second, the cost of that patience is slower incorporation of information that is not yet public and a18price path that becomes predictable; the variance ratio doubles at full penetration. Third, coordination19among LLM agents is high but not, in itself, harmful: the agents agree on the direction of the public20value, prompt and model diversity reduce their agreement and speed up absorption at a liquidity cost, and21a communication channel or memory raises agreement and pricing error by amounts that are precisely22estimated but small.2324The honest summary of the pre-registered hypotheses is that H2 (synchrony), H4 (diversity) and H525(communication) are supported in sign, H1 (informativeness) is supported for dynamics and rejected for26levels, H3 (instability) is rejected and its leverage corollary reversed, and H6 (safeguards) is not27testable at the shock sizes used. The stylized model that organizes the paper is a model of28market-order traders with correlated errors; the data call for a model of limit-order traders with a29common anchor. Writing that model, giving the LLM agents leverage of their own, running the Claude cells30that the design contains but a rejected key prevented, and raising the seed count to the level that31detects half-standard-deviation effects are the next steps, and the replication package makes each of32them a configuration change.3334Nothing here should be read as evidence about live markets. The claim is narrower and, we think, more35useful: when the share of trading decisions taken by one language model is raised exogenously in a36market where everything else is held fixed, the first-order effect is not herding into bubbles but a37deep, calm and slow market---and the slowness, not the calm, is where the next question lies.38