Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenAI Progress

progress.openai.com

Post reply on HN
Print view View original

361–370 of 372 posts

Re: OpenAI Progress

#361

Post by OnlineGladiator »

glenstein wrote: post #360
OnlineGladiator wrote: post #354

Earlier quoted context omitted.

So your goal here is to say the same thing over and over again and hope I eventually give the affirmation you so desperately need? You've already declared that you're right multiple times. Nobody cares but you. https://xkcd.com/386/ You might want to develop a sense of humor. You'll enjoy life more.

My goal is to invite you to think critically about the specific caveats in the comment you are replying to instead of ignoring those caveats. They said that generally speaking using thinking mode on non niche topics they can get reliable answers, and invited anyone who disagreed with it to offer examples where it fails to perform as expected, a constructive structure for counter examples in case anyone disagreed. You…

As I've said many times before, I am aware of everything you have said. I just don't care. You seem to be really upset that someone on the internet disagrees with you. And from my perspective, you are the one that has no self-awareness and is completely missing the point. You don't even understand the conversation we're having and yet you're constantly condescending.

I'm sure if you keep repeating yourself though I'll change my mind.

Re: OpenAI Progress

#362

Post by mh- »

platevoltage wrote: post #345
mh- wrote: post #244

Earlier quoted context omitted.

As you identified, not paying for it is a big part of the issue. Running these things is expensive , and they're just not serving the same experience to non-paying users. One could argue this is a bad idea on their part, letting people get a bad taste of an inferior product. And I wouldn't disagree, but I don't know what a sustainable alternative approach is.

I would have no issue if the free version of ChatGPT told me straight up “You gotta pay for links and sources”. It doesn’t do that.

100% agree with that, as I alluded to in my last sentence. And that honestly seems like it might be a good product strategy in the short term.

Re: OpenAI Progress

#363

Post by hauntsaninja »

Dilettante_ wrote: post #246
fastball wrote: post #234

Why did they call GPT-3 "text-davicini-001" in this comparison? Like, I know that the latter is a specific checkpoint in the GPT-3 "family", but a layman doesn't and it hardly seems worth the confusion for the marginal additional precision.

Thanks for noting that, as I am a layman who didn't know.

text-davinci-001 is just not GPT-3 in any real sense

(I work at OpenAI, I helped build this page and helped train text-davinci-001)

Re: OpenAI Progress

#364

Post by glenstein »

OnlineGladiator wrote: post #361
glenstein wrote: post #360

Earlier quoted context omitted.

My goal is to invite you to think critically about the specific caveats in the comment you are replying to instead of ignoring those caveats. They said that generally speaking using thinking mode on non niche topics they can get reliable answers, and invited anyone who disagreed with it to offer examples where it fails to perform as expected, a constructive structure for counter examples in case anyone disagreed. You…

As I've said many times before, I am aware of everything you have said. I just don't care. You seem to be really upset that someone on the internet disagrees with you. And from my perspective, you are the one that has no self-awareness and is completely missing the point. You don't even understand the conversation we're having and yet you're constantly condescending. I'm sure if you keep repeating yourself though I'l…

Simianwords said: "use GPT 5 with thinking and search disabled and get it to give you inaccurate facts for non niche, non deep topics" and noted that mistakes were possible, but rare.

JustExAWS replied with an example of getting Python code wrong and suggested it was a counter example. Simianwords correctly noted that their comment originally said thinking mode for factual answers on non-niche topics and posted a link that got the python answer right with thinking enabled.

That's when you entered, suggesting that Simian was "missing" the point that GPT (not distinguishing thinking or regular mode), was "not always right". But they had already acknowledged multiple times that it was not always right. They said the accuracy was "high enough", noted that LLMs get coding wrong, and reiterating that their challenge was specifically about thinking mode.

You, again without acknowledging the criteria they had noted previously, insisted this was cherry picking, missing the point that they were actually being consistent from the beginning, inviting anyone to give an example showing otherwise. At no point between then and here have you demonstrated an awareness of this criteria despite your protestations to the contrary.

Instead of paying attention to any of the details you're insulting me and retreating into irritated resentment.

Re: OpenAI Progress

#365

Post by OnlineGladiator »

glenstein wrote: post #364
OnlineGladiator wrote: post #361

Earlier quoted context omitted.

As I've said many times before, I am aware of everything you have said. I just don't care. You seem to be really upset that someone on the internet disagrees with you. And from my perspective, you are the one that has no self-awareness and is completely missing the point. You don't even understand the conversation we're having and yet you're constantly condescending. I'm sure if you keep repeating yourself though I'l…

Simianwords said: "use GPT 5 with thinking and search disabled and get it to give you inaccurate facts for non niche, non deep topics" and noted that mistakes were possible, but rare. JustExAWS replied with an example of getting Python code wrong and suggested it was a counter example. Simianwords correctly noted that their comment originally said thinking mode for factual answers on non-niche topics and posted a lin…

Thank you for repeating yourself again. It's really hammering home the point. Please, continue.

Re: OpenAI Progress

#366

Post by ninetyninenine »

OnlineGladiator wrote: post #365
glenstein wrote: post #364

Earlier quoted context omitted.

Simianwords said: "use GPT 5 with thinking and search disabled and get it to give you inaccurate facts for non niche, non deep topics" and noted that mistakes were possible, but rare. JustExAWS replied with an example of getting Python code wrong and suggested it was a counter example. Simianwords correctly noted that their comment originally said thinking mode for factual answers on non-niche topics and posted a lin…

Thank you for repeating yourself again. It's really hammering home the point. Please, continue.

He’s right. you’re not saying anything that actually furthers the conversation or counters any of his points.

Re: OpenAI Progress

#367

Post by muzani »

andai wrote: post #199
muzani wrote: post #198

Earlier quoted context omitted.

ChatGPT was a proper product, but as an engine, GPT-3 (davinci-001) has been my favorite all the way until 4.1 or so. It's absolutely raw and they didn't even guardrail it. 3.5 was like Jenny from customer service. davinci-001 was like Jenny the dreamer trying to make ends meet by scriptwriting, who was constantly flagged for racist opinions. Both of these had an IQ of around 70 or so, so the customer service trainin…

davinci-002 is still available, and pretty close.

Sorry, I disagree. There's a reason it's 002. They cleaned up all the racist, etc, opinions from 001.

I do think GPT-4.1 onwards has a lot of personality. It's able to pretend to go into this mindset and go back out, which works fine. If I wanted to talk to actual racists, there's plenty out there. But I just want the spicy flavor in my AI because it's a little bland otherwise.

Re: OpenAI Progress

#368

Post by muzani »

robrenaud wrote: post #222
muzani wrote: post #195

Earlier quoted context omitted.

I don't have a source for this (there's probably no sources from anything back then) but anecdotally, someone at an AI/ML talk said they just added more data and quality went up. Doubling the data doubled the quality. With other breakthroughs, people saw diminishing gains. It's sort of why Sam back then tweeted that he expected the amount of intelligence to double every N years. I have the feeling they kept on this u…

The input size to output quality mapping is not linear. This is why we are in the regime of "build nuclear power plants to power datacenters". Fixed size improvements in loss require exponential increases in parameters/compute/data.

Oh yeah, I forgot about this. Something about the quality going up linearly as the data doubled - this was even in OpenAI's documentation for fine tuning.

Re: OpenAI Progress

#369

Post by OnlineGladiator »

ninetyninenine wrote: post #366
OnlineGladiator wrote: post #365

Earlier quoted context omitted.

Thank you for repeating yourself again. It's really hammering home the point. Please, continue.

He’s right. you’re not saying anything that actually furthers the conversation or counters any of his points.

I was never trying to counter any of his points. I think we just disagree on what makes for good conversation.

Re: OpenAI Progress

#370

Post by glenstein »

OnlineGladiator wrote: post #365
glenstein wrote: post #364

Earlier quoted context omitted.

Simianwords said: "use GPT 5 with thinking and search disabled and get it to give you inaccurate facts for non niche, non deep topics" and noted that mistakes were possible, but rare. JustExAWS replied with an example of getting Python code wrong and suggested it was a counter example. Simianwords correctly noted that their comment originally said thinking mode for factual answers on non-niche topics and posted a lin…

Thank you for repeating yourself again. It's really hammering home the point. Please, continue.

You keep insisting that you understand, but the upshot of understanding would have been acknowledging the specifics which you haven't done despite clearly having plenty of time and energy to reply.

Instead you've tried everything from saying I need to "get a sense of humor", to character attacks, to insisting without specific explanation that I "don't understand", to declaring that you "don't care", to declaring that no amount of information will make you acknowledge the inaccuracy of your own comments.

So you haven't succeeded in changing the subject of the conversation, except in the sense of turning it into a tutorial about why you can't make wrong into right with character attacks or declarations about how much you don't care.

Post reply on HN