Even a hyper-rational mind with infinite computation capacity will draw the wrong conclusions if it is fed incorrect observations. LLMs aren't magic, they're pa

Even a hyper-rational mind with infinite computation capacity will draw the wrong conclusions if it is fed incorrect observations. LLMs aren't magic, they're pattern-matching engines. Even if the relationship between input and output is opaque, that doesn't mean they're omniscient.