LLM agents run a town's economy in new paper
What happens if you agents to run a town's economy? This super interesting paper provides some insi...
This paper provides concrete data on how LLM agents perform in a complex economic simulation, showing that even with a 12x boost, money flow can stall and wealth rankings are misleadingly stable early on.
A new paper simulates 100 LLM agents managing a town's economy for 26 simulated weeks. The agents earn wages, run businesses, and set prices on Pokhara Lakeside geography across 91 runs and 2.44M decisions. A 12x tourist shock raises business revenue 4.62x, but wages only move 1.03x and 0.3% of 3,981 menu items are repriced. A cash transfer shows the same pattern, with 96.7% of it unspent 311 steps later. Swapping the underlying LLM changed every outcome, while deleting memory had no effect. Wealth rankings look frozen at 2 weeks but only start to move by week 12.
What happens if you agents to run a town's economy? This super interesting paper provides some insi...
What happens if you agents to run a town's economy? This super interesting paper provides some insights: They put 100 LLM agents in charge of a town economy for 26 simulated weeks, and they find that money stops moving. The agents earn wages, run businesses and set prices on real Pokhara Lakeside geography, across 91 runs and 2.44M decisions. A 12x tourist shock raises business revenue 4.62x. Wages move 1.03x, and only 0.3% of 3,981 menu items are ever repriced. A cash transfer shows the same pattern. 96.7% of it is still unspent 311 steps later. Here is the interesting part for anyone building agent simulations. Swapping the underlying LLM changed every outcome they measured. Deleting the agents' memory changed none of them detectably. Results at the usual 1 to 2 week horizon also mislead. Wealth rankings look frozen at 2 weeks and only start to move by week 12. Paper: arxiv.org/abs/2609.11108 Chat with Paper: academy.dair.ai/papers/but-how… 💬 9 🔄 1 ❤️ 26 👀 3330 📊 13 ⚡