News for hackers. Stuff that matters.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by andrewstuart on Sunday October 19, 2025 @08:53AM (#45633777)
Humans hallucinating about AI.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by JKCalhoun on Sunday October 19, 2025 @09:25AM (#45633959)
"OpenAI Researcher Hallucinates GPT-5 Math Breakthrough" could be a headline from The Onion.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by reaperducer on Sunday October 19, 2025 @10:08AM (#45634251)
"OpenAI Researcher Hallucinates GPT-5 Math Breakthrough" could be a headline from The Onion.

Off topic, but I saw The Onion on sale in the magazine rack of Barnes and Noble last month.

For those who miss when it was a free rag in sidewalk newsstands, and don't want to pony up for a full subscription, this is an option.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by antegamisou on Sunday October 19, 2025 @09:42AM (#45634062)
Seriously those headlines are getting DailyMail sensationalism levels of ridiculous.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by nicce on Sunday October 19, 2025 @09:54AM (#45634145)
I the old world we would just use the word bullshit.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by alkyon on Sunday October 19, 2025 @10:28AM (#45634392)
They started believing the very lies they invented.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by moffkalast on Sunday October 19, 2025 @10:32AM (#45634411)
"The truth is usually just an excuse for a lack of imagination."
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by pera on Sunday October 19, 2025 @09:29AM (#45633985)
Heh stockholders are not hallucinating: They know very well what they are doing.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by skeeter2020 on Sunday October 19, 2025 @10:10AM (#45634273)
retail investors? no way. The fever-dream may continue for a while but eventually it will end. Meanwhile we don't even know our full exposure to AI. It's going to be ugly and beyond burying gold in my backyard I can't even figure out how to hedge against this monster.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by pera on Monday October 20, 2025 @07:19AM (#45642642)
Yeah no I didn't mean retail investors, OpenAI is not publicly traded, but yeah I do share your concern...
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by Palmik on Sunday October 19, 2025 @09:52AM (#45634131)
More like humans hallucinating about humans hallucinating about AI, see here: https://news.ycombinator.com/item?id=45634120
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by MattGaiser on Sunday October 19, 2025 @09:21AM (#45633932)
Humans "hallucinate" in the AI way constantly, which is why I don't see them as a barrier to LLMs replacing humans in many contexts. It really isn't unusual for a human to make stuff up or be unaware of stuff.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by zeknife on Sunday October 19, 2025 @09:30AM (#45633993)
A human being informed of a mistake will usually be able to resolve it and learn something in the process, whereas an LLM is more likely to spiral into nonsense
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by MattGaiser on Sunday October 19, 2025 @09:51AM (#45634125)
You must know people without egos. Humans are better at correcting their mistakes, but far worse at admitting them.

But yes, as an edge case handler humans still have an edge.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by topaz0 on Sunday October 19, 2025 @09:58AM (#45634183)
LLMs by contrast love to admit their mistakes and self-flagellate, and then go on to not correct them. Seems like a worse tradeoff.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by skeeter2020 on Sunday October 19, 2025 @10:13AM (#45634300)
Not when your goal is to create ASI: Artificial Sycophant Intelligence
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by thaumasiotes on Sunday October 19, 2025 @10:04AM (#45634218)
It's true that the big public-facing chatbots love to admit to mistakes.

It's not obvious to me that they're better at admitting their mistakes. Part of being good at admitting mistakes is recognizing when you haven't made one. That humans tend to lean too far in that direction shouldn't suggest that the right amount of that behavior is... less than zero.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by tonyhart7 on Sunday October 19, 2025 @10:01AM (#45634193)
and this is why LLM is getting cooked

they feed an internet data into that shit, they basically "told" LLM to behave because surprise surprise, human sometimes can be more nasty

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by alimw on Sunday October 19, 2025 @09:36AM (#45634029)
You must know better humans than I do.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by skeeter2020 on Sunday October 19, 2025 @10:12AM (#45634292)
Do you think the OpenAI human, when informed of their "oopsie" replied "You're right, there is existing evidence that this problem has already been solved. Blah Blah Blah ... and that's why our new model has made a huge breakthrough against previously unsolved math problems!"
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by zeroonetwothree on Sunday October 19, 2025 @09:52AM (#45634134)
Humans are a bit better at knowing which things are important and doing more research. Also better at being honest when directly pressed. And infinitely better at learning from errors.

(Yes, not everyone, but we do have some mechanisms to judge or encourage)

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by pas on Sunday October 19, 2025 @09:28AM (#45633979)
it's the same thing with self-driving, if you can make it safer than a good human driver it's enough. but the bar is pretty low with driving (as evidenced by the hundreds of thousands of collisions and deaths and permanent disabilities each year). and rather high in scientific publishing.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by lapcat on Sunday October 19, 2025 @09:53AM (#45634141)
> Humans "hallucinate" in the AI way constantly

This claim is ambiguous. The use of the word "Humans" here obscures rather than clarifies the issue. Individual humans typically do not "hallucinate" constantly, especially not on the job. Any individual human who is as bad at their job as an LLM should indeed be replaced, by a more competent individual human, not by an equally incompetent LLM. This was true long before LLMs were invented.

In the movie "Bill and Ted's Excellent Adventure," the titular characters attempt to write a history report by asking questions of random strangers in a convenience store parking lot. This of course is ridiculous and more a reflection of the extreme laziness of Bill and Ted than anything else. Today, the lazy Bill and Ted would ask ChatGPT instead. It's equally ridiculous to defend the wild inaccuracy and hallucinations of LLMs by comparing them to average humans. It's not the job of humans to answer random questions on any subject.

Human subject matter experts are not perfect, but they’re much better than average and don’t hallucinate on their subjects. They also have accountability and paper trails, can be individually discounted for gross misconduct, unlike LLMs.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by topaz0 on Sunday October 19, 2025 @12:27PM (#45635487)
> Humans "hallucinate" in the AI way constantly

This is more and more clearly false. Humans get things wrong certainly, but the manner in which they get things wrong is just not comparable to how the LLMs get things wrong, beyond the most superficial comparison.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by random9749832 on Sunday October 19, 2025 @09:21AM (#45633933)
Best case: Hallucination

Worst case (more probable): Lying

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by MPSimmons on Sunday October 19, 2025 @09:28AM (#45633981)
Hanlon's Razor
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by random9749832 on Sunday October 19, 2025 @10:09AM (#45634262)
They are expanding into the adult market because they are running out of ideas. I think common sense is enough to decide what is what here.
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by forgetfulness on Sunday October 19, 2025 @09:31AM (#45634011)
Lying is a stupid way of selling something and making money
Re:OpenAI researcher announced GPT-5 math breakthrough that ... by reaperducer on Sunday October 19, 2025 @10:09AM (#45634258)
Lying is a stupid way of selling something and making money

Works for Elon.

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by rixed on Sunday October 19, 2025 @09:29AM (#45633990)
These days AI just obsequiously praise whatever stupid ideas the human throw at them, which encourage humans into hallucinating breakthroughs.

But it's only a matter of time before AI gets better at prompt engineering.

/s?

Re:OpenAI researcher announced GPT-5 math breakthrough that ... by Kiboneu on Sunday October 19, 2025 @05:17PM (#45638073)
No no, openai is actually secretly run by AI.