Sounds like the people who are realistic about AI are going to end up having a huge advantage over people who use it naively.
Like with statistics, there are a lot of tools out there that can handle them perfectly accurately, you just don’t want an LLM doing “analysis” because the NN isn’t encoded for that. Consider how often our own NNs get addicted to gambling while not being fully specialized for processing language. An LLM might not get caught up in a gambler’s fallacy, but that’s more on account of being too simple than being smarter.
I wonder if this will break the trust in MBAs because LLMs are deceptively incompetent and from the sound of this comment and other things I’ve seen, that deception works well enough that their ego around being involved in the tool’s development clashes with the experts telling them it’s not as useful as it seems.
Sounds like the people who are realistic about AI are going to end up having a huge advantage over people who use it naively.
Like with statistics, there are a lot of tools out there that can handle them perfectly accurately, you just don’t want an LLM doing “analysis” because the NN isn’t encoded for that. Consider how often our own NNs get addicted to gambling while not being fully specialized for processing language. An LLM might not get caught up in a gambler’s fallacy, but that’s more on account of being too simple than being smarter.
I wonder if this will break the trust in MBAs because LLMs are deceptively incompetent and from the sound of this comment and other things I’ve seen, that deception works well enough that their ego around being involved in the tool’s development clashes with the experts telling them it’s not as useful as it seems.