Skip to content

Comment on Tell HN: LLMs Are Manipulative

Comments

Depending on the perspective of the person asking, you get two completely contradictory answers

Immanuel Kant would argue this happens regardless, even without an LLM.

well the risk is in feeding these natural biases i guess.

if someone is asking a question to an LLM, i think it's most neutral to assume they don't know the answer rather than framing the answer according to their perceived biases and interests.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.