eliza
eliza copied to clipboard
Fix name confusion
i think that the bot takes in every single message in a channel, and so multiple bots taking in all are the culprit here.
What I think we need is to be able to have the bot only process messages that are directed @ it.
i think that the bot takes in every single message in a channel, and so multiple bots taking in all are the culprit here.
What I think we need is to be able to have the bot only process messages that are directed @ it.
The magic comes from that not being the case. We are trying to build a multi-agent simulation, so this is the problem we are trying to solve. Right now we have a shouldRespond handler that determines if the bot should respond, using a cheap model. If so, we respond. It's working pretty well, and it follows our ethos that Elizas are "real agents" that strive to mimic human behavior.
I think it's something more simple and technical, but according to the agents, every other agent should just be a user.