Using AgentSpeak(L) as a DSL for event loop / control flow for LLM-based Autonomous Agents, like SuperAGI/BabyAGI/AutoGPT/MetaGPT/BeeBot/etc

[aindilis]

Hi Niklas,

Thank you for releasing python-agentspeak!

I've been watching LLM-based Autonomous Agents take off. One thing that bothers me is the control flow for the agents seems quite simple. Whereas AgentSpeak(L) has a more sophisticated event loop and data representation.

I was wondering what the status of python-agentspeak is? Do you have plans to resume work on it? Also, I am curious what you might think about my use case, whether python-agentspeak is a good framework for controlling LLM-agents? I.e. a skeleton upon which to put some LLM muscle.

Unfortunately I'm just starting out with Python myself, so I cannot really contribute.

I think if we can get python-agentspeak<->LLM-based-automous-agent, that will be checkmate.

Cheers,
Andrew

[javipalanca]

Hi Andrew,
Unfortunately @niklasf is not answering messages lately, he is probably very busy with other things. I am interested in continuing with the project and possibly we will soon publish a new version (although I would have loved to have Niklas' supervision).

Regarding the use of LLMs, it is a very interesting approach and maybe we will soon examine it in my line of research. But for the moment we are dealing with other issues and I don't think it will be immediate....