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

GPT-Neo – Building a GPT-3-sized model, open source and free

eleuther.ai

Post reply on HN
Print view View original

31–40 of 260 posts

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#31

Post by phreeza »

taneq wrote: post #14
anticristi wrote: post #3

I love the initiative, but I'm starting to get scared of what a post-GPT-3 world will look like. We are already struggling to distinguish fake news from real ones, automated customer request replies from genuine replies, etc. How will I know that I have a conversation with a real human in the future? On the other side, the prospect of having an oracle that answers all trivia, fixes spelling and grammar, and allows hu…

The fake news thing is a real problem (and may become worse under GPT3 but certainly exists already). As for the others - to quote Westworld, "if you can't tell the difference, does it really matter?"

Genuine question, why is this a problem? Sure, someone may be able to generate thousands of real-sounding fake news articles, but it's not like they will also be able to flood the New York Times with these articles. How do you worry you will be exposed to these articles?

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#32

Post by querez »

spiderfarmer wrote: post #8
Erlich_Bachman wrote: post #4

It's a shame that it has turned out to be necessary to externally re-make and re-train a model that has come out of company called `OPEN`AI. Wasn't one of the founding principles of it that all of the research would be available to the public? Isn't that the premise on which the initial funding was secured? Best of luck to Eleuther.

I don't really care if OpenAI offers commercial licenses as long as the underlying research is truly open. This way alternative options will become available eventually.

Arguably openAI is one of the most closed industry AI labs (among those that are still participating in the research community), on par only with deep mind (though deepmind at least publishes way more). Funnily enough, FAIR and Google Brain have a vastly better track record wrt. publishing not only papers but also code and models.

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#33

Post by anticristi »

hntrader wrote: post #29
turing_complete wrote: post #9

Earlier quoted context omitted.

Don't read news. Go to original sources and scientific papers. If you really want to understand something, a news website should only be your starting point to look for keywords. That is true today as it will be "post-GPT-3".

That's what people should do, and that's what you and I will do, but many won't, especially the less educated (no condescension intended). They'll buy into the increased proliferation of fake info. It's because of these people that I think the concerns are valid.

Honestly, I consider myself fairly educated (I have a PhD in CS), but if the topic at hand is sufficiently far from my core competence, then reading the scientific article won't help. I keep reading about p-value hacking, subtle ways of biasing research, etc., and I realize that, to validate a scientific article, you have to be a domain expert and constantly keep up-to-date with today's best standards. Given the increasing number of domains to be an expert in, I fail to see how any single human can achieve that without going insane. :D

I mean, Pfizer could dump their clinical trial reports at me, and I would probably be unable to compute their vaccine's efficiency, let alone find any flaws.

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#35

Post by anticristi »

normanmatrix wrote: post #15
anticristi wrote: post #3

I love the initiative, but I'm starting to get scared of what a post-GPT-3 world will look like. We are already struggling to distinguish fake news from real ones, automated customer request replies from genuine replies, etc. How will I know that I have a conversation with a real human in the future? On the other side, the prospect of having an oracle that answers all trivia, fixes spelling and grammar, and allows hu…

You will not. Welcome to the scary generative future.

I was hoping for a "yes, we can" attitude here. :D

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#36

Post by wizzwizz4 »

turing_complete wrote: post #9
anticristi wrote: post #3

I love the initiative, but I'm starting to get scared of what a post-GPT-3 world will look like. We are already struggling to distinguish fake news from real ones, automated customer request replies from genuine replies, etc. How will I know that I have a conversation with a real human in the future? On the other side, the prospect of having an oracle that answers all trivia, fixes spelling and grammar, and allows hu…

Don't read news. Go to original sources and scientific papers. If you really want to understand something, a news website should only be your starting point to look for keywords. That is true today as it will be "post-GPT-3".

> Go to original sources and scientific papers.

Given how much bunk “science” (and I'm talking things completely transparent to someone remotely competent in the field) gets published, especially in psychology, it's difficult to do even that.

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#37

Post by Pyramus »

danielscrubs wrote: post #17
Erlich_Bachman wrote: post #6

Earlier quoted context omitted.

Photoshop has existed for decades. Is it really that big of a problem for photo news?

Touch ups where done before photoshop but now it’s ALWAYS done. The issues this has created in society might have a bigger emotional impact than we give it credit for. Regarding photo news there has been quite a lot of scandals to the point that I’d guess the touchups is more or less accepted.

I conducted a workshop in media compentency for teenage girls and one of the key learnings was that every image of a female subject they encounter in printed media (this was before Instagram) has been retouched.

To hammer the point home I let them retouch a picture by themselves to see what is possible even for a completely untrained manipulator.

It was eye-opening - one of the things that should absolutely be taught in school but isn't.

Re: GPT-Neo – Building a GPT-3-sized model, open source and free

#39

Post by mhuffman »

Erlich_Bachman wrote: post #4

It's a shame that it has turned out to be necessary to externally re-make and re-train a model that has come out of company called `OPEN`AI. Wasn't one of the founding principles of it that all of the research would be available to the public? Isn't that the premise on which the initial funding was secured? Best of luck to Eleuther.

But I was told GPT-3 was too powerful for mere mortal hands (unless you have an account!) and that it would be used for hate speech and to bring about skynet.

How will this project avoid those terrible outcomes?

Post reply on HN