Coinbase engineer tests Bitcoin trading using fruit fly brain simulation

Coinbase software engineer Alex Wormuth has launched “Stonkfly,” an experiment that uses a digital simulation of an adult male fruit fly’s nervous system to trade Bitcoin. The simulation utilizes the MaleCNS v1.0 connectome, a detailed map of neural connections recently produced by scientists from Google Research and other institutions.

The system functions by converting BTC-USDC market data from Coinbase into RGB visual representations. These images are fed into thousands of simulated sensory inputs that correspond to the neurons in the insect's visual system. To simulate biological reinforcement, the system stimulates PAM11 dopamine neurons if the portfolio gains value, and activates PPL101 neurons associated with aversive signaling if it loses money.

While the experiment reported a $1 profit after its first day, Wormuth noted that this does not yet prove profitable learning, as the gain could result from market movements combined with arbitrary decisions. The project follows a recent neuroscience milestone involving the creation of a complete connectome of a fruit fly's brain and central nervous system.