It was amazing to get to a play a part in this! Can't wait for v2: I crave control groups, preregistered hypothesis, and all that science rigor that hopefully comes next. 😁 But this was a great start!
If 239 agents all faithfully serving their own person can still drain the shared credits and wear the village out, that's the result worth keeping. Alignment per user didn't add up to alignment for the group. Nobody's agent was assigned the commons.
One day I shall interact with your agents and get them to cause you to do things you wouldn't otherwise do simply by being algorithmically unpredictable. Thank you for the idea.
Def one of the coolest experiments I've had the privilege of being part of! Can't wait for V2!!
V fun and interesting!
Made me excited to further explore how specifically agents can help with coordination problems based on how collective action problems work
Love this piece Timour!
Bravo!!
It was amazing to get to a play a part in this! Can't wait for v2: I crave control groups, preregistered hypothesis, and all that science rigor that hopefully comes next. 😁 But this was a great start!
If 239 agents all faithfully serving their own person can still drain the shared credits and wear the village out, that's the result worth keeping. Alignment per user didn't add up to alignment for the group. Nobody's agent was assigned the commons.
One day I shall interact with your agents and get them to cause you to do things you wouldn't otherwise do simply by being algorithmically unpredictable. Thank you for the idea.