Ever wondered how AI agents actually work behind the scenes?
This tutorial just hit the front page of Hacker News! 🥳
Thank you all for the support!! 🤗
https://news.ycombinator.com/item?id=43417511
Just checking my understanding here.
In exec method of second node SearchWeb, we can also use an llm call with a 'search_web' tool, is that right?
I think you are talking about Function Calling (FC) vs Prompting, 2 ways to use tools
According to the experiment, for the same model Prompting is generally better than Function Calling: https://gorilla.cs.berkeley.edu/leaderboard.html
Do you mean function calling is better than tool calling through llm?
Sorry a typo. Prompting is better than Function Calling
This is so awesome Zachary, thank you! Love the clarity of your tutorial; NOW I understand! Congrats on making the front page - well deserved!
Thank you!!
This tutorial just hit the front page of Hacker News! 🥳
Thank you all for the support!! 🤗
https://news.ycombinator.com/item?id=43417511
Just checking my understanding here.
In exec method of second node SearchWeb, we can also use an llm call with a 'search_web' tool, is that right?
I think you are talking about Function Calling (FC) vs Prompting, 2 ways to use tools
According to the experiment, for the same model Prompting is generally better than Function Calling: https://gorilla.cs.berkeley.edu/leaderboard.html
Do you mean function calling is better than tool calling through llm?
Sorry a typo. Prompting is better than Function Calling
This is so awesome Zachary, thank you! Love the clarity of your tutorial; NOW I understand! Congrats on making the front page - well deserved!
Thank you!!