Fuck AI

hideshow 7 child comments

[–] belit_deg@lemmy.world 23 points 1 year ago*

...And that people take the bait and anthropomorphize it, believing it is "reasoning" and "thinking".

It seems like people want to believe it because it makes the world more exciting and sci-fi for them. Even people who don't find gpt personally useful, get carried away when talking about the geopolitical race to develop agi first.

And I sort of understand why, because the alternative (and I think real explanation) is so depressing - namely we are wasting all this money, energy and attention on fools' gold.

parent

[–] Gradually_Adjusting@lemmy.world 8 points 1 year ago (4 children)

If these things could not be induced to confidently lie, they would not be the target of billions of dollars of investments.

hideshow 4 child comments

parent

[–] Lemminary@lemmy.world 3 points 1 year ago* (3 children)

I sincerely don't think this is true, but it's a nice narrative that fits well with one of Lemmy's. It'd still be worth the same or more if it hallucinated to a minimum because it would better match one of its ideal business applications: replacing human labor at a fraction of the cost. Unfortunately, this is only a convenient side effect for many who stand to benefit from creating propaganda and false information in bulk.

hideshow 3 child comments

parent

[–] Gradually_Adjusting@lemmy.world 9 points 1 year ago (2 children)

I prefer to live in a world where it is possible to say something oversimplified and ridiculous, and people just laugh and don't feel like it deserves correction. That's practically everything I ever feel like saying.

hideshow 2 child comments

parent

[–] Lemminary@lemmy.world 2 points 1 year ago (1 child)

Sorry! I know the feeling and I wasn't meaning to be a jerk about it by ackshually'ing you. I mostly replied because I think Lemmy could do better at criticizing LLMs in general.

parent

hideshow 1 child comment

[–] Gradually_Adjusting@lemmy.world 3 points 1 year ago

I agree actually. It's hard to get everyone on the same page about much of anything, and the topic is both very large and changing rapidly. It's a lot less work to just make vague sarky remarks about venture capitalists.

parent

[–] pulsewidth@lemmy.world 1 point 1 year ago

And yet there are still people in the thread claiming that 'oh ChatGPTs knowledge data cuts off at the end of 2024, this prompt is using ChatGPT wrong' completely missing your point.

If ChatGPT doesn't know something it just lies about it, all while being passed off as doctorate-level intelligence.

Inb4 defenders 'an AI can't lie it just asserts falsehoods as truth because it's having a scary dream/hallucination' as if semantics will save the day.