The Age of Agents: Can multiple LLM agents improve outputs? (opens in new tab)

(sudoapps.substack.com)

2 pointssudoapps2y ago3 comments

3 comments

This was a quick experiment on creating multiple LLM (GPT) agents with their own objective to see if outputs can be improved.

As in, a consensus network? IMO, yes. Is that scary? Yes.

Very similar to a consensus network. That is probably the future of LLM agents.

j / k navigate · click thread line to collapse

This was a quick experiment on creating multiple LLM (GPT) agents with their own objective to see if outputs can be improved.

As in, a consensus network? IMO, yes. Is that scary? Yes.

Very similar to a consensus network. That is probably the future of LLM agents.

j / k navigate · click thread line to collapse