3 ms·

The problem with Artificial "Intelligence" is that it really has no intelligence at all. Intelligence requires understanding, and AI doesn't understand either t

The problem with Artificial "Intelligence" is that it really has no intelligence at all. Intelligence requires understanding, and AI doesn't understand either the data fed into it or the responses it gives. Yet because these tools output confident, plausible-sounding answers with a professional tone (which may even be correct a majority of the time), they give a strong illusion of being reliable. What will be the result of the current push of GPT AI into the mainstream? If people start relying on it for things like summarizing articles and scientific papers, how many wrong conclusions will be reached as a result? God help us if doctors and engineers start making critical decisions based on generative AI answers.

danans 4y ago
> What will be the result of the current push of GPT AI into the mainstream? If people start relying on it for things like summarizing articles and scientific papers, how many wrong conclusions will be reached as a result? God help us if doctors and engineers start making critical decisions based on generative AI answers. On the other hand, it may end up completely undermining its own credibility, and put a new premium on human sourced information. I can see 100% human-sourced being a sort of premium label on information in the way that we use "pesticide-free" or "locally-sourced" labels today.
- gptgpp 4y ago
  Nice! This would make for a super fun sci-fi... The poors that need medicine get put in front of an LLM that gets it right most of the time, if they're lucky enough to have a common issue / symptomatic presentation. Hey, when you're poor, you can't afford a one-shot solution! You gotta put up with a many-shot technique. Meanwhile the rich people get an actual doctor that can use sophisticated research and medical imaging. Kindly human staff with impeccable empathy and individualized consideration -- the sort of thing only money can buy.

danans 4y ago

> What will be the result of the current push of GPT AI into the mainstream? If people start relying on it for things like summarizing articles and scientific papers, how many wrong conclusions will be reached as a result? God help us if doctors and engineers start making critical decisions based on generative AI answers. On the other hand, it may end up completely undermining its own credibility, and put a new premium on human sourced information. I can see 100% human-sourced being a sort of premium label on information in the way that we use "pesticide-free" or "locally-sourced" labels today.

gptgpp 4y ago
Nice! This would make for a super fun sci-fi... The poors that need medicine get put in front of an LLM that gets it right most of the time, if they're lucky enough to have a common issue / symptomatic presentation. Hey, when you're poor, you can't afford a one-shot solution! You gotta put up with a many-shot technique. Meanwhile the rich people get an actual doctor that can use sophisticated research and medical imaging. Kindly human staff with impeccable empathy and individualized consideration -- the sort of thing only money can buy.