3 ms·
If you want AI to be useful it will eventually encounter untrusted content, such as via web search. I think things like web search should probably be run on a d
by PokestarFan 2mo ago
If you want AI to be useful it will eventually encounter untrusted content, such as via web search. I think things like web search should probably be run on a different sandboxed AI whose task is to write a summary that is then ingested by the main agent, similar to how existing sandboxing already works, but this would diminish the usefulness quite a bit.
- savanaly 2mo ago>I think things like web search should probably be run on a different sandboxed AI whose task is to write a summary that is then ingested by the main agent, similar to how existing sandboxing already works, but this would diminish the usefulness quite a bit. It also wouldn't work. You would simply mindjack the outer AI and have it mindjack the inner AI in turn with its summary. Nesting AIs can't fix the malicious input problem.
- santadays 2mo agoJust pass that through a third llm.
- InsideOutSanta 2mo agoCorrect, that has been prophesied by scripture: "Thou shalt have three layers of LLMs, no more, no less. Three shall be the number thou shalt have, and the number of the counting of the LLM layers shall be three."
- s_Hogg 2mo agoFour is right out
- dbetteridge 2mo ago> Once the number three, being the third number, be reached, then lobbest thou thy Holy LLM of Antioch towards thy task, who, being naughty in My sight, shall snuff it.