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

GPT-4 details leaked?

threadreaderapp.com

Post reply on HN
Print view View original

501–510 of 648 posts

Re: GPT-4 details leaked?

#501

Post by notahacker »

ly3xqhl8g9 wrote: post #399
notahacker wrote: post #348

Earlier quoted context omitted.

Strangely, it is not news to me that exploitation is a normal day-to-day part of any industry . (This is why I have not made any statements to that effect, never mind done anything as ludicrous as post articles about unionised screenwriters seeking to negotiate a higher pay rate as evidence that "most" of them earned "nothing") And this is also why I am not advocating a copyright-free world in which HuffPo has the ri…

Hope at least you are getting paid and you don't deploy such ridiculous amounts of bad faith for free. My reply from above "Most of them already don't get paid, hence the strike" was specifically in the context of the streaming platforms not paying residuals in the same manner the studios do, hence the reference to the article. What I find most funny is that you don't even know or care about alternative solutions, yo…

What I find most funny is that you're implying there are alternative solutions that you do know and care about.

And yet instead of addressing my actual objection - that screenwriters seeking to get paid more for co-creating IP would be unlikely to see abolition of IP protection as a solution - by articulating those alternatives and how it would allow them to get paid more, you chose to assert they don't get paid.[1]

I mean, I'm not the one advocating the radical change here, even though actually I don't "assume that the copyright laws is for the best in the best of all possible worlds". So it's not really incumbent on me to answer my own objections by imagining screenwriter-satisfying solutions involving the abolition of copyright. If you actually had one and were able to advocate it with as much zeal as you have defended the claim "most of them already don't get paid" this might be a more interesting discussion.

[1]doesn't really matter if you were doing so "specifically in the context of streaming platform residuals" since whether or not they get residuals from a particular media type is irrelevant to the fact screenwriters are not in favour of proposals which would entail them losing both the residuals they do get and their job

Re: GPT-4 details leaked?

#502

Post by ben_w »

mdp2021 wrote: post #419
ben_w wrote: post #276

Earlier quoted context omitted.

> And safely here means non disruptive to established businesses Why would OpenAI care about that? No, it means safety, as in not giving out dangerous answers that get people killed.

> as in not giving out dangerous answers that get people killed Literally taken, that is quite close to impossible. It was news days ago of somebody who committed suicide after having an interaction with a bot about nuclear risks or similar. To avoid that, the bot would have to be a high-ranking professional psychologist with an explicit purpose not to trigger destructive reactions. And that would fail the nature of…

> To avoid that, the bot would have to be a high-ranking professional psychologist with an explicit purpose not to trigger destructive reactions.

Sounds fairly dooable, TBH, given what else they're doing reasonably well at.

> And that would fail the nature of a "consultant", which is something "under best effort to speak the truth" - incompatible direction with "something reassuring".

I don't think they're as incompatible as you appear to.

Re: GPT-4 details leaked?

#503

Post by logicchains »

lossolo wrote: post #442
Vecr wrote: post #123

Earlier quoted context omitted.

You can run that model (Wizard-30) on a computer with 64 gigabytes of RAM (or smaller, I don't know how tight you can cut it). You obviously want fast RAM and a good CPU, but you don't need a GPU.

You can also travel on a bike from NY to LA.

Just imagine if Boeing lobbied the government to pass a law banning you biking from NY to LA.

Re: GPT-4 details leaked?

#504

Post by avereveard »

mdp2021 wrote: post #419
ben_w wrote: post #276

Earlier quoted context omitted.

> And safely here means non disruptive to established businesses Why would OpenAI care about that? No, it means safety, as in not giving out dangerous answers that get people killed.

> as in not giving out dangerous answers that get people killed Literally taken, that is quite close to impossible. It was news days ago of somebody who committed suicide after having an interaction with a bot about nuclear risks or similar. To avoid that, the bot would have to be a high-ranking professional psychologist with an explicit purpose not to trigger destructive reactions. And that would fail the nature of…

if that is the bar to pass for content for public consumption, we're dangerously close to book burning, again.

Re: GPT-4 details leaked?

#505

Post by ben_w »

ilaksh wrote: post #426
ben_w wrote: post #236

Earlier quoted context omitted.

The public position (as opposed to the rumour mills) is that they're not working on a 5, and don't intend to at least until they can figure out how to do it safely.

I believe that the superalignment thing will basically be GPT-5.

I think it's a necessary precursor to any future model. Would you want to spend ten times as much (or whatever) red-teaming the next increase in capabilities? And what would make you more confident in advance that the red team would spot everything?

Re: GPT-4 details leaked?

#506

Post by londons_explore »

droopyEyelids wrote: post #473
makestuff wrote: post #387

Earlier quoted context omitted.

The moat is there until someone releases (or leaks) comparable training data.

And that gets more difficult every day, as previously accesible sources of data turn off their api. though google may have something up its sleeve with the corpus of google books! I have been wondering if openAI secretly pulled in scihub or zlibrary to neutralize that potential advantage.

Are there any stats on how many words are in google books, vs how many words are on the open web?

My feeling is that the web has a lot more on it than the total of all libraries - simply because anyone can start a blog, but publishing a book requires quite some commitment.

Re: GPT-4 details leaked?

#507

Post by snowycat »

Vecr wrote: post #123
Llamamoe wrote: post #103

Earlier quoted context omitted.

How low can you get the memory and computational power requirements that way?

You can run that model (Wizard-30) on a computer with 64 gigabytes of RAM (or smaller, I don't know how tight you can cut it). You obviously want fast RAM and a good CPU, but you don't need a GPU.

I am running 30b llama models (4 bit quantized using llama.cpp) on 32 gb of ram and no GPU. I get around 2 tokens/second.

Re: GPT-4 details leaked?

#508

Post by javajosh »

YeGoblynQueenne wrote: post #229
PostOnce wrote: post #19

"Open" AI, a charity to benefit us all by pushing and publishing the frontier of scientific knowledge. Nevermind, fuckers, actually it's just to take your jobs and make a few VCs richer. We'll keep the science a secret and try to pressure the government into making it illegal for you to compete with us. https://github.com/ggerganov/llama.cpp https://github.com/openlm-research/open_llama https://huggingface.co/TheBlok…

>> We'll keep the science a secret and try to pressure the government into making it illegal for you to compete with us. Just to be clear, there's no science being kept secret because there is no science being done. OpenAI's is a feat of engineering, borne aloft by a huge budget supporting a large team whose expertise lies in tuning neural net systems, and not in doing science. Machine learning, as it is practiced to…

Obligatory xkcd: https://xkcd.com/1838/

Re: GPT-4 details leaked?

#509

Post by javajosh »

gitfan86 wrote: post #451
jncfhnb wrote: post #428

Earlier quoted context omitted.

Sure. But remember also that the standard “science” is largely a broken model. Academia sucks. It’s honestly bad. It’s overflowing with papers that are misleading, irrelevant, or fraudulent due to a mix of poor stats knowledge, bad incentives, and a failure to organize effectively. While people praise the scientific method, the majority of achievements we attribute to “science” are not derived from guess and check gr…

Right. In historical context the scientific method was a huge improvement over the previous method of understanding the world, which was mostly religion. But today, academia is a victim of organizational and political capture, making it less competitive for talent.

I remain convinced that the correct model for science is a) serious hobbyists and b) patronage for the really talented. It would weed out the careerists, at least. You just can't make a "science factory" and expect to produce discoveries like cogs. It's stupid to even try.

Re: GPT-4 details leaked?

#510

Post by stOneskull »

juliensalinas wrote: post #100
kyledrake wrote: post #65

Earlier quoted context omitted.

Unfortunately I've found the current OSS models to be vastly inferior to the OpenAI models. Would love to see someone actually get close to what they can do with GPT-3.5/4, except capable of running on commodity GPUs. What's the most impressive open model so far?

LLaMA 30B or 60B can be very impressive when correctly prompted. Deploying the 60B version is a challenge though and you might need to apply 4-bit quantization with something like https://github.com/PanQiWei/AutoGPTQ or https://github.com/qwopqwop200/GPTQ-for-LLaMa . Then you can improve the inference speed by using https://github.com/turboderp/exllama . If you prefer to use an "instruct" model à la ChatGPT (i.e. tha…

> The interesting thing with these Uncensored models is that they don't constantly answer that they cannot help you (which is what ChatGPT and GPT-4 are doing more and more)

that's great to hear. that political correctness in gpt is annoying.

Post reply on HN