In a current experiment, Anthropic created a confidential market the place AI brokers characterize each patrons and sellers, conducting actual transactions for actual items and actual cash.
The corporate confirmed that the take a look at, known as Mission Deal, was solely a “pilot experiment with a self-selected participant pool” of 69 Anthropic staff who got a price range of $100 (paid for in present playing cards) to purchase issues from co-workers.
Nonetheless, Anthropic mentioned it was “shocked by how nicely the challenge offers labored,” with 186 transactions totaling greater than $4,000.
The corporate mentioned it really operated 4 completely different marketplaces with completely different fashions. One was “actual” (all had been represented by the corporate’s cutting-edge fashions and had been really permitted for buying and selling after testing), and the opposite three had been for analysis functions.
Apparently, customers get “objectively higher outcomes” when they’re represented by a extra subtle mannequin, Anthropic mentioned. Nonetheless, customers don’t appear to concentrate on this disparity, growing the potential of a “distinction within the high quality of brokers” wherein “individuals on the shedding aspect is probably not conscious that their scenario is getting worse.”
It additionally seems that the preliminary directions given to the brokers didn’t have an effect on the probability of a sale or the negotiated worth.
