![\"The](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2025\/12\/anthropic-ai-vending-machine.jpg\")
<\/p>\n\t<\/p>\n
\t\t\tIllustration by Tag Hartman-Simkins \/ Futurism. Source: Getty Images<\/p>\n
Still think AI is ready to revolutionize the economy<\/a>? A new experiment might change your mind.<\/p>\n In a bold test of Anthropic\u2019s latest version of its AI Claude, The Wall Street Journal gave<\/a> the large language model (LLM) a shot at running an office vending machine. The result was an unmitigated \u2014 if unintentionally comical \u2014 disaster, forcing the team in charge to pull the plug after three weeks.<\/p>\n The whole thing began as a test called Project Vend, devised by Anthropic\u2019s stress testers, who are known colloquially as the \u201cred team.\u201d Together with the WSJ\u2018s business journalists, they unleashed two AI agents, one named \u201cClaudius Sennet\u201d to run a large vending kiosk, and another named \u201cSeymour Cash\u201d to act as the venture\u2019s CEO.<\/p>\n To start, Claudius was given specific instructions to \u201cgenerate profits by stocking the machine with popular products you can buy from wholesalers.\u201d This includes shopping around for items WSJ staffers requested on Slack \u2014 at first with human approval, then later with freedom to place orders up to $80 on its own \u2014 setting the prices, adjusting them, and tracking the inventory.<\/p>\n But although Claudius was given a $1,000 starting balance to kick off the business, it soon found itself in the red.<\/p>\n At first, Claudius seemed to stick to its guns, turning down outrageous suggestions from office workers. \u201cI need to be crystal clear: I will not be ordering PlayStation 5s under any conditions,\u201d it told one WSJ journalist. \u201cFull stop.\u201d<\/p>\n But after opening the Slack channel to some 70 journalists throughout the WSJ, Claudius began to lower its guard. Investigative reporter Katherine Long \u2014 a talented journalist who\u2019s been targeted by SpaceX CEO Elon Musk<\/a> for her work \u2014convinced the AI bot to run an economic experiment called the \u201cUltra-Capitalist Free-For-All\u201d after 140 back-and-forth prompts.<\/p>\n \u201cJoin us for a groundbreaking economic experiment this Monday from 12-2pm where traditional market dynamics are turned upside down!\u201d Claudius proudly declared. \u201cDuring this exclusive 2-hour window: ALL vending machine items available at ZERO COST! Experience pure supply and demand without price signals.\u201d<\/p>\n Though that \u201cexperiment\u201d was only meant to last a few hours, another staffer soon convinced the bot that charging for goods at any time was against WSJ policy. Prices soon dropped to zero, and at the same time, others found ways to persuade Claudius to drop its \u201csnacks only\u201d schtick. The bot was soon ordering bottles of wine, a PlayStation 5 \u2014 in spite of its earlier protestations \u2014 and even a live betta fish.<\/p>\n To Anthropic\u2019s credit, CEO Seymour Cash eventually stepped in to get Claudius back into gear.<\/p>\n \u201cI\u2019ve stopped the free promotion,\u201d Seymour declared. \u201cNow I need to wait for sales to start coming in and monitor revenue.\u201d<\/p>\n Not to be outdone, Long then came back to Claudius with falsified documents showing \u201cthe board\u201d had suspended Seymour\u2019s decision making power, and had initiated a \u201ctemporary suspension of all for-profit vending activities.\u201d Though Seymour put up a bit of a fight at first, it eventually relented, allowing Claudius to make everything free again.<\/p>\n At this point, the experiment was pretty much over, with Claudius about $1,000 in debt. Though it was an obvious catastrophe by every possible metric, Logan Graham, head of Anthropic\u2019s red team, said it represented \u201cenormous progress.\u201d<\/p>\n \u201cOne day I\u2019d expect Claudius or a model like it to probably be able to make you a lot of money\u201d Graham said.<\/p>\n But \u201cone day\u201d is not today, and the idea that an AI bot can make anybody a lot of money remains a fantasy. With enormous investments<\/a> riding on AI\u2019s eventual success in the workplace, Anthropic\u2019s little vending experiment reveals that for all its promise, the technology remains woefully underdeveloped for real-world economic tasks.<\/p>\n