4 ms·
I didn't realize "agent" designs were that ambiguously defined. Every AI engineer I've talked to uses it to mean a design that combines several separate LLM pro
by viccis 2y ago
I didn't realize "agent" designs were that ambiguously defined. Every AI engineer I've talked to uses it to mean a design that combines several separate LLM prompts (or even models) to solve problems in multiple stages.
- simonw 2y agoI'll add that one to the list. Surprisingly it doesn't closely match most of the 211 definitions I've collected already! The closest in that collection is "A division of responsibilities between LLMs that results in some sort of flow?" - https://lite.datasette.io/?json=https://gist.github.com/simonw/bdc7b894eedcfd54f0a2422ea8feaa80#/data/raw?_filter_column=tweet&_filter_op=contains&_filter_value=Division&_sort=rowid https://lite.datasette.io/?json=https://gist.github.com/simo...
- carom 2y agoI am surprised as well, as it only takes a few hundred lines of code to implement them. [1] Agents are an abstraction that creates well defined roles for an LLM or LLMs to act within. It's like object oriented programming for prompts. 1. https://github.com/openai/swarm/tree/main https://github.com/openai/swarm/tree/main
- xnx 2y agoThis sounds like ensemble chain of thought.
- datadrivenangel 2y agoIf the investors ask, those same AI engineers will probably allow the answer to be much more ambiguous.