In a current experiment, Anthropic created a categorized market the place AI brokers represented each patrons and sellers, hanging actual offers for actual items and actual cash.
The company admitted this take a look at — which it referred to as Challenge Deal — was solely “a pilot experiment with a self-selected participant pool” of 69 Anthropic staff who got a funds of $100 (paid out by way of reward playing cards) to purchase stuff from their coworkers.
Nonetheless, Anthropic stated it was “struck by how nicely Challenge Deal labored,” with 186 offers made, totaling greater than $4,000 in worth.
The corporate stated it truly ran 4 separate marketplaces with completely different fashions — one which was “actual” (the place everybody was represented by the corporate’s most-advanced mannequin, and with offers truly honored after the experiment) and one other three for research.
Apparently, when customers are represented by extra superior fashions, they get “objectively higher outcomes,” Anthropic stated. However customers didn’t appear to note the disparity, elevating the potential for “‘agent high quality’ gaps” the place “folks on the shedding finish may not notice they’re worse off.”
Additionally, the preliminary directions given to the brokers didn’t seem to have an effect on sale probability or the negotiated costs.
