When you typed this response, you were acting as a probabilistic, predictive chat model. You predicted the most likely effective sequence of words to convey ideas. You did this using very different circuitry, but the underlying strategy was the same.
Comment on We need to stop pretending AI is intelligent
aesthelete@lemmy.world 3 weeks agoHow is outputting things based on things it has learned any different to what humans do?
Humans are not probabilistic, predictive chat models. If you think reasoning is taking a series of inputs, and then echoing the most common of those as output then you mustn’t reason very well or very often.
FourWaveforms@lemm.ee 3 weeks ago
aesthelete@lemmy.world 3 weeks ago
I wasn’t, and that wasn’t my process at all. Go touch grass.
FourWaveforms@lemm.ee 2 weeks ago
I would rather smoke it than merely touch it, brother sir
stephen01king@lemmy.zip 2 weeks ago
Then, unfortunately, you’re even less self-aware than the average LLM chatbot.
aesthelete@lemmy.world 2 weeks ago
Dude chatbots lie about their “internal reasoning process” because they don’t really have one.
Writing is an offshoot of verbal language, which during construction for people often has more to do with sound and personal style than the popularity of words. It’s not uncommon to bump into individuals that have a near singular personal grammar and vocabulary and that speak and write completely differently with a distinct style of their own.
But you know, you have to leave your house, touch grass, and actually listen to some people speak that aren’t talking heads on television in order to discover that truth.
NotASharkInAManSuit@lemmy.world 2 weeks ago
By this logic we never came up with anything new ever, which is easily disproved if you take two seconds and simply look at the world around you. We made all of this from nothing and it wasn’t a probabilistic response.
Your lack of creativity is not a universal, people create new things all of the time, and you simply cannot program ingenuity or inspiration.
chunes@lemmy.world 3 weeks ago
Do you think most people reason well?
The answer is why AI is so convincing.
aesthelete@lemmy.world 3 weeks ago
I think people are easily fooled. I mean look at the president.
kibiz0r@midwest.social 3 weeks ago
Image
DancingBear@midwest.social 3 weeks ago
This is great