Maximilian Kasy

@maxkasy.bsky.social

Econ prof at Oxford. Machine learning, politics, econometrics, inequality, random reading recs. maxkasy.github.io/home/

This creates opportunities for local / democratic control: - Take open-weight foundation models - Fine tune for local / democratically chosen objectives using RL. Here some useful reading recs: 2/5

Thread: Key developments in AI / LLMs 1. reinforcement learning (RL, from human feedback or verifiable rewards), 2. open weight models; some based on distillation of commercial models, not far behind commercial alternatives. 1/5

The Emerging Market for Intelligence: How Firms Buy and Sell AI

(Summer 2026) - We describe the emerging business-to-business market for large language model (LLM) inference and document key empirical patterns in its supply, pricing, and dynamics, using data from…

aeaweb.org

Thread: Key developments in AI / LLMs 1. reinforcement learning (RL, from human feedback or verifiable rewards), 2. open weight models; some based on distillation of commercial models, not far behind commercial alternatives. 1/5

The Emerging Market for Intelligence: How Firms Buy and Sell AI

(Summer 2026) - We describe the emerging business-to-business market for large language model (LLM) inference and document key empirical patterns in its supply, pricing, and dynamics, using data from…

aeaweb.org

So, naive question about these campuses that pay OpenAI $13mil/yr. Why don’t they pour that money into a cluster running the strongest available open-weight model, with API and interactive options. And if the answer is “overhead,” why aren’t we collaborating?