You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
First, kudos on building this innovative AI agent framework!
I'd like to propose an experimental concept for discussion: What if we enable recursive AI capabilities by adding LLM/VLM models - and potentially owl itself - to the tool pool?
The text was updated successfully, but these errors were encountered:
Cool! That's exactly what I am thinking about. We were actually already explored some applications. For example, when an web agent are viewing a webpage, it may want to use code executor to process some task, then it will call code agent. Feel free to contact me for further discussions via: [email protected]
Well, instead of using predefined agents like code agent, can we let the system generate prompts dynamically, making the LLM/VLM in the tool pool act as needed roles, that would be more flexible. The system uses LLV/VLM just like human beings.
First, kudos on building this innovative AI agent framework!
I'd like to propose an experimental concept for discussion: What if we enable recursive AI capabilities by adding LLM/VLM models - and potentially owl itself - to the tool pool?
The text was updated successfully, but these errors were encountered: