www.volksverpetzer.de/unterstutzen...

What could be the utility of LLMs?

But aren’t we already using in many realms of our lives many other statistical models to predict future and taking decisions based on it? Like weather models, financial models, aren’t elections being very influenced by opinion polls? Do we question this statistic elements in our lives enough?

So LLMs are talking like we talk without knowing what they are saying. Because they can learn a set of rules but they can’t know, experience and develop morality?

Are you saying we shouldn’t take LLMs answers for good, right or real because the intelligence in LLMs is not what we understand under it for humans but it is more a good enough sentence built on statistic probability?

Is this translated to the LLMs question, bias and the way machine learning takes place and how answers are being decoded?

Hello Dave, thanks for your thread. Are you computer engineer? I would like to write a couple of questions under your thread!