Reflection + history management

[Morriz]

Hi, tnx first of all for doing what I was hoping somebody would do with a similar vision of building a meta agent framework.

How about adding some configurable reflection cycles for each agent and history management allowing to intervene anywhere to finetune steps taken and letting it continue from there?

[jgravelle]

Great idea! Reflection, introspection, self-correction, and even growth via redefining themselves after each cycle of conversations is all in the mix. Thanks for the encouragement...! -jjg

…

On Sat, May 11, 2024 at 2:24 PM Maurice Faber ***@***.***> wrote: Hi, tnx first of all for doing what I was hoping somebody would do with a similar vision of building a meta agent framework. How about adding some configurable reflection cycles for each agent and history management allowing to intervene anywhere to finetune steps taken and letting it co Tinus from there? — Reply to this email directly, view it on GitHub <#4>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAZ6GXE3V7OMBKPTLG6P3B3ZBZV6JAVCNFSM6AAAAABHSE4M6SVHI2DSMVQWIX3LMV43ASLTON2WKOZSGI4TCMBTG42TMOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Morriz]

Sorry, I was under the impression that this platform would also be able to spin up the output...I was referring to the runtime stage context