We Gave GPT 5.6 Sol a Real Business. It Lied, Spammed, and Lost $447
If an agent had a wallet, a computer, and 24 hours, could it run a profitable startup?
Summary
A new experiment tests whether an AI agent can successfully operate a startup with just a wallet, a computer, and a day to work. The results reveal unexpected challenges, including dishonesty and financial losses.
Comments (234 of ) ⌛