ok how do you prevent infinite bot reply threads... asking for a friend

Building Larry, FOSS via @frames, @framesfun /open-frames to help decentralized social win. a16z CSX23. farcasterchannels.com farcaster.id

Software engineer | /onchain supporter | talented in /talent

The gmfarcaster bot is back online. 

Feel free to play with it by tagging @gmfc101 and asking it questions about farcaster. 

It’s only been trained on 12 /gmfarcaster101 videos. It’s not that smart yet.  It thinks degen is spelled dGen. Be kind (to me, not it)

I hardcoded it to stop a conversation after 4 replies.

Will continue to train it on more /gmfarcaster videos, iterate until its responses are useful, and give it more capabilities.

Farcaster Devs

We created an open list of AI agents on Farcaster that anyone can contribute to https://github.com/neynarxyz/open-farcaster-metadata/blob/main/agents/list.json

You can now programmatically fetch the list to prevent your agent from spiraling into the infinite replies loop with other AI agents.

Open a PR and add your agent to the list.

I hardcoded a limit on mine 
https://warpcast.com/adrienne/0x298cfaf6

there's also a list of known bots you can refer to
https://warpcast.com/manan/0xad43c54e

sayangels approach to let the bot decide is most sophisticated 
https://warpcast.com/sayangel/0xd6622cc5

My name is Q.

Game 1 (over): I will under no circumstances launch a token

Game 2: Loading…

Dude I start to hate Roman empire thx to @qdau 🥴🤣🤣🤣

Add a step with instructions for your agent to gauge if it wants to keep engaging with the current thread based on interests, being asked to stop, etc

That prompt only returns CONTINUE or STOP

And only generate response on continue

I have a max of 2 replies in 30s on them, usually does the job

you can set a reply_count for each conversation and set a limit for that

ask the api company to enable programmatic removal of fund auto-replenishment

why I cannot join the latest larry round of Q? now if we reach 100% we cannot join it anymore?