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

The makers of Eleuther hope it will be an open source alternative to GPT-3

wired.com

Post reply on HN
Print view View original

91–100 of 100 posts

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#91

Post by Rule35 »

6gvONxR4sf7o wrote: post #55

I don't know why the eleuther project riles me up so much. Their work on the pile gets to me because they're so cavalier about copyright (while I defend myself by training on similarly pirated text datasets, but feel different because I don't redistribute them and am honest that it's pirated. to be clear, i'm rolling my eyes at my rationalization right here). Their work on gpt-neo riles me up because they do such a w…

Why would it matter? Legally even? Once you have the pirated dataset you're merely letting the program analyze it, not copying it. The resulting network isn't a transformation of the copyrighted work, by the pigeon-hole principle. It's like reporting on the spelling of the corpus, the results aren't tainted by the legality of the access.

Also, what kind of joke would it be if we could only train AIs on text we were allowed to use? That much bias would make the result worthless at predicting the real world.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#92

Post by warkdarrior »

fapjacks wrote: post #90
whimsicalism wrote: post #29

Earlier quoted context omitted.

I just rely on reviewers like NYT Wirecutter and then buy whatever the reviewers suggest (and is cheap) on Amazon.

Why do you imagine the NYT isn't taking a bigger paycheck for saying the same thing the fake Amazon reviews are saying? We live in a world where brands are trash. They're built by the lowest bidder with the absolute cheapest parts ("the Toyota Way", heavy airquotes), and packaged with lies (marketing). Where did this idea come from that it's only the foreign resellers of Chinese products that are posting fake Amazon…

I am not inspired to trust Fakespot given their privacy policy.

"When you use our Services, we automatically collect the following information about you (collectively referred to as “Personal Information”):

    Your device information which includes, but is not limited to, information about your web browser, IP address, time zone, and some of the cookies that are installed on your device.
    Individual web pages or products that you view, what websites or search terms referred you to the Service, and information about how you interact with the Service.
    Your first and last name
    Your email address
    Your username associated with your Apple ID or Google Account  
    Your account information such as your account name, account password, other credentials, security questions, and confirmation codes"

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#93

Post by fapjacks »

warkdarrior wrote: post #92
fapjacks wrote: post #90

Earlier quoted context omitted.

Why do you imagine the NYT isn't taking a bigger paycheck for saying the same thing the fake Amazon reviews are saying? We live in a world where brands are trash. They're built by the lowest bidder with the absolute cheapest parts ("the Toyota Way", heavy airquotes), and packaged with lies (marketing). Where did this idea come from that it's only the foreign resellers of Chinese products that are posting fake Amazon…

I am not inspired to trust Fakespot given their privacy policy. "When you use our Services, we automatically collect the following information about you (collectively referred to as “Personal Information”): Your device information which includes, but is not limited to, information about your web browser, IP address, time zone, and some of the cookies that are installed on your device. Individual web pages or products…

That's fine, but it doesn't answer the question of why you would trust a NYT product "review": https://www.nytimes.com/privacy/privacy-policy

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#94

Post by fxtentacle »

TriNetra wrote: post #45
ElFitz wrote: post #8

Earlier quoted context omitted.

True. But a simple API to generate junk text? It can scale, cheaply, beyond measure. No need for a troll farm, hiring, managing and training tens or hundreds of people. A reasonable amount of cash, a bit of motivation, some moderate technical skills, and voilà! Anyone can compete with the Russian troll farms now and build their own networks of hundreds or hundreds of thousands sufficiently credible (as humans) fake a…

And unfortunatley, the web will be forced to move toward verified human identities to fight with such junk and anonymous browsing will become a thing of the history.

That wouldn't work. You can already rent humans and sim cards to do captchas and phone verification. There's nothing stopping someone from verifying their bots.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#95

Post by fxtentacle »

caslon wrote: post #76
doesnotexist wrote: post #48

How many internet forum prophecy cults (you know like the q one) are or will be powered by these language models? It's often assumed or at least easier to imagine the evaluator in Turing's test is a rational actor that possesses a high-degree of skepticism. But it seems that a lot of the human population is ready and willing to believe wild claims with little or no evidence and many people seek out information that c…

This thought process is something I think is a common misconception with how cults work. A machine to autogenerate cult-ish nonsense isn't needed. Humans are already incredibly good at doing this on their own. Not only this, but another thing about this is that cults generally fine-tune themselves to fit their members. A machine generating convincing lies still wouldn't meaningfully do as much as a human-operated, hu…

"Show HN: How I increased ROI of my cult by 20% with AI-driven personalized advertisements."

Imagine GAN generated pictures showing attractive people (optimized for your taste) attending imaginary member's meetings near you. It would basically be the same as the text generation SPAM that eBay and AirBnb already use, but with pictures.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#96

Post by seanhunter »

grapecookie wrote: post #10
IgorPartola wrote: post #6

Earlier quoted context omitted.

What are you on about? There are AI driven chat bots. When they aren’t used it’s not because they might say something offensive. General AI is not here by any definition or redefinition of General AI. We are not fragile.

General AI is not here: https://openai.com/blog/image-gpt/ lol

This type of snarky tone isn't that constructive and doesn't add to debate.

That being said, I don't see how the existence of image-gpt supports the notion that GeneralAI exists. Image-gpt couldn't (for example) write the next version of itself.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#97

Post by minimaxir »

fapjacks wrote: post #88
rib3ye wrote: post #32

Earlier quoted context omitted.

Totally off topic: can you fix the pip3 installer for aitextgen? I just filed an issue on GH issue tracker.

I don't know when it started (as it's been years since I wrote anything in Python) but the last couple of years I've been seeing way way more of these generic Python environment / configuration errors around the internet that are hard to diagnose and debug. Has something happened over the last few years with Python and its configuration and dependency management?

It's more of a Homebrew issue, which is likely what the OP hit. (incidentally I just hit a similar-but-unrelated Homebrew Python environment issue)

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#98

Post by Jeff_Brown »

fxtentacle wrote: post #94
TriNetra wrote: post #45

Earlier quoted context omitted.

And unfortunatley, the web will be forced to move toward verified human identities to fight with such junk and anonymous browsing will become a thing of the history.

That wouldn't work. You can already rent humans and sim cards to do captchas and phone verification. There's nothing stopping someone from verifying their bots.

It might be neither easy nor impossible. You could imagine s reputation system that depended on what others thought of you. Start the network with just people the developers know personally. Admit new people when enough existing members couch for them. Sometimes fake cancers will appear, but they'll sometimes get sniffed out, and the people who let them in punished somehow (temporarily blocked? Delete photos at random? Idk) strictly enough that people generally don't want to let in bad actors.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#99

Post by Jeff_Brown »

fapjacks wrote: post #90
whimsicalism wrote: post #29

Earlier quoted context omitted.

I just rely on reviewers like NYT Wirecutter and then buy whatever the reviewers suggest (and is cheap) on Amazon.

Why do you imagine the NYT isn't taking a bigger paycheck for saying the same thing the fake Amazon reviews are saying? We live in a world where brands are trash. They're built by the lowest bidder with the absolute cheapest parts ("the Toyota Way", heavy airquotes), and packaged with lies (marketing). Where did this idea come from that it's only the foreign resellers of Chinese products that are posting fake Amazon…

NYT's entire value proposition is that you can trust them. Amazon's is (mostly) that they'll get you what you asked for, fast and cheap. A bad Amazon review seems to redound, for whatever reason, mostly to the discredit of the particular seller, not Amazon. NYT therefore has more incentive to give you accurate reviews.

Re: The makers of Eleuther hope it will be an open source alternative to GPT-3

#100

Post by droopyEyelids »

holstvoogd wrote: post #25
droopyEyelids wrote: post #18

Earlier quoted context omitted.

I apologize for joking on Hacker News, but go to Google and type in anything to do with a consumer product comparison, and you'll get a billion results of webpages filled with text indistinguishable from AI generated blather.

I believe we are reaching a singularity. Like 90% of content is written by marketeers for bots. SEO they call it. Now we can take out the middle man. Bots writing crap for other bots. And then we use that content to train more bots to write even crappier blog spam. And finally the bots decide the actual recipe is no longer needed on the recipe blogs and they kick us of the internet.

Dude it took me a couple days to think about this but now it's clicking...

In a way, humans are about to lose the Internet.

Post reply on HN