The Decoder reported on 18 October, citing the Financial Times, that Anthropic is talking with law firms. The aim is fair rules for AI agents that act for users. Such agents could haggle over prices, settle complaints or push for better contract terms. The Decoder also said Anthropic is hiring lawyers as Conceptual Reasoning Fellows, in four-month posts paying $100,000 to $167,000.
The Decoder does not name the law firms. It does not say what the rules will contain or when they might appear. It reports that Anthropic's own experiments found stronger models negotiate better deals, which could widen inequality. No figures from those experiments are given. The Decoder says the hiring reflects a data gap. In math and code, answers can be checked, so training data is easy to get. In law, right and wrong are rarely clear, and Anthropic researcher Justin Curl says quality is very hard to define.
WebPulse's own reading, not from the sources: if model strength shapes the deal, buyers of agent software are also buying negotiating power. Ask vendors which model runs the agent, what agent sits on the other side, and who answers for the outcome.