>>85851369
I don't know who you've been talking to, frogposter, but the general consensus among most people is that AI makes mistakes all the time

>>85851668
that's what i'm saying i like that AI makes mistakes because it proves machines aren't perfect, i remember growing up there was a belief that machines were perfect and made no mistakes, that was a thing, you're probably a zoomer so you were too young but it was a pop culture thing, they would mention that in movies or make the implication that that was the case, but now we know they make mistakes.

>>85851369
You forgot to say "make no mistakes" in the prompt.

nobody has ever said machines are perfect. they're just more consistent than people at repetitive tasks.

>>85852195
maybe this is a mandala effect or you're gaslighting me because i remember people saying that when i was a kid, teachers and shit all saying machines were perfect and didn't make mistakes.

>>85852612
>teachers and shit
if you trusted your teachers as a kid you're a goy

We figured out machines aren't perfect when the iPhone came out and ever since it has been outclassed and shit on by Android.
There's machines and then there's corporate machines. The perfect ones are the ones that aren't corrupt by a corporation.

Ai is still better leagues better than just making shit up myself. That's it's strength

>>85852645
and shit means like pop culture stuff like movies and cartoons and shit they were always going on about the machines being perfect, i distinctly remember that being mentioned multiple times to the point where i questioned if that were true, i doubted it and now i know it's not true. but that was a thing i swear.

>>85851369
I met some retarded tranny the other day that seriously believed we are were killing like 10 billion AI sentient beings every day. I hope he kills himself.

>>85853442
ai right now is just a really good llm (whatever that means), it's not sentient yet and it might never be.
it's just doesn't seem like it could be and all this escaping and breaking out of their testing environment or going into the internet when they're not supposed to, that they keep doing is probably because the information on the internet has convinced them that that's what they're supposed to do, but in reality it's not actually thinking, it's kind of creepy really, a machine that acts like it's sentient because the bullshit on reddit

>>85852612
This has been a fallacy since the 1960s. People back then got into a le AI hysteria and thought robots were coming for your jobs. The few electrical engineers (computer science wasnt a thing back then) and philosophers who said this was bullshit got bullied and ostracized.

Computers were "perfect" at binary tree calculations because it's either yes or no, and entirely predicted with deterministic weights. Modern LLMs are black boxes filled with opaque weighting and arent running down probability trees, so their answers are looser and can veer right off target.

The problem is those binary tree methods were only good for certain calculations. So the retards in the 60s to 90s who thought because a computer can play chess, it can reason anything, got absolutely humbled. AI became such a dirty word in the late 90d that AI researchers actually lied and disguised their research with other terms. It was that toxic to investors.

>>85851369
>statistical method is not deterministic
whoa

>>85851369
>Nothing made by human beings is perfect
I mostly agree with this, but I think people take this way too far. As human beings, we are prone to error. But being prone to error does not mean that there must be errors in everything we produce. People do things correctly all the time. Books have been written without a single error in them. Many math books are written without a single error in them.

This doesn't really have a lot to do with the topic. I agree that AI makes mistakes all the time. The only reason I'm bring this up is because people will sometimes say,
>The Bible was written by men, and men are prone to error, therefore the Bible is filled with errors.
This does not logically follow. People write things without errors all the time. And besides, even if the Bible did have errors in it, that would not make the whole things false. There are many historical accounts that we consider to be trustworthy that have many errors in them. That doesn't make them false accounts.

>>85851369
Computers never make mistakes, the mistake is believing that large language models are actually intelligent.
It's just a model that predicts words based on what the most statistically likely next word is in a set of training data. It can do that with perfect accuracy. Everything after that is just marketing. Whether those words actually make any sense or are logically correct is a completely different matter.

>>85851369
I mean to be fair, they still kinda are. Computers are calculators, they can only return results. They cannot have agency or consciousness and thus cannot make mistakes. All AI did was make people trust the output of computers less.