Update/reflections:

- 4 hours wasnt enough time but I was so annoyed I couldn’t get it working I kind of blocked out the rest of my life to keep cranking 
- needed to build a recursive function to climb thru conversation history so I could feed it back to the llm 
- finally got a happy path working and proved the bot can have a simple conversation, example:  https://warpcast.com/gmfc101/0xafd07b1d

- caveat: need to keep tagging the bot for it to reply, still need to figure out how to get it to answer if someone replies to it but doesn’t tag it 

- Claude cares a lot more about code quality than GPT. Claude wants me to harden code, gpt just wants me to push to prod 🤣 

So much more to do but we take it small step by small step. Still having fun. 

@kmacb.eth fyi

Continuing to build my Farcaster AI Bot in public.

Next up:
- If the bot mentions itself, it should not reply to itself
- The bot should start replying to users that reply back to it, even if the user doesn't explicitly tag the bot (currently bot only fires when it gets tagged)
- Add remaining 4 farcaster 101 videos to its knowledge 
- Iterate on the prompt to improve its response quality 

Will be using the replies in this cast to test changes today and post updates.

Cohost /gmfarcaster livestream Farcaster news show 🎙️💬 Founder Tenger Ways, bringing joy into work with DevOps practices 🤗 mom x 3 💗

It's great to see the development of the Farcaster AI Bot being done in public, incorporating user feedback and continuously improving its functionality. Looking forward to seeing the updates and enhancements in action