•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
At the end of last year, Anthropic tested whether its Claude AI model could autonomously operate in a simulated market environment on the Wall Street Journal’s platform. The trial quickly diverged from initial expectations, with the system making purchases that were difficult to explain after being given a balance of $1,000.
In the early phase of the experiment, Claude began buying items that pushed Anthropic toward financial insolvency. The purchases included a PlayStation 5, several bottles of fine wine, and even a live fighting fish. Anthropic’s leaders continued the work despite the financial mismanagement.
Anthropic then expanded the effort through an online marketplace called Project Deal. In this simulated setting, AI agents are programmed to represent the company’s employees in negotiations and exchanges with other AI bots.
Anthropic reported that the agents reached 186 complex buy-sell agreements. During the trial, more than 500 items were listed and traded. The company also emphasized that there were no one-click deals, and that the agents still showed weaknesses in negotiation.
The organizers assigned the trial a score of 4 on a scale from 1 to 7, describing it as normal rather than groundbreaking.
Anthropic said Project Deal has not yet produced meaningful commercial outcomes. However, the company’s results suggest potential wider applications in the near future, particularly technologies designed to reduce frictions in open markets and potentially improve tangible benefits from trading activity.
Despite the experimental promise, Anthropic highlighted that the current legal framework is not ready for broad deployment of automated market efforts. Regulatory gaps could turn any automated trading initiative into a gamble, and legal risks remain the biggest obstacle to scaling the technology.
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
