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

Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

chatllama.baseten.co

Post reply on HN
Print view View original

61–70 of 226 posts

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#61

Post by eob »

huggingmouth wrote: post #54
O__________O wrote: post #51

>> Not trying to shill anything; just want to build and share cool stuff. Any commercial use of Facebook’s LLaMa remains a violation of the software’s terms of use to my understanding; not saying your use is a violation, but that as it relates to anything related to LLaMa there no related commercial use.

This is wrong. Terms of service only apply if you agree to them. They do do not magically apply if you download the weights from someone else. Both the weights and the fine-tuned data were generated without human creativity and thus not eligible for copyright protection under law.

Has a court decided that yet? That’s huge if true.

Someone once told me a curation was impossible to copyright for similar reasons.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#62

Post by codingdave »

huggingmouth wrote: post #54
O__________O wrote: post #51

>> Not trying to shill anything; just want to build and share cool stuff. Any commercial use of Facebook’s LLaMa remains a violation of the software’s terms of use to my understanding; not saying your use is a violation, but that as it relates to anything related to LLaMa there no related commercial use.

This is wrong. Terms of service only apply if you agree to them. They do do not magically apply if you download the weights from someone else. Both the weights and the fine-tuned data were generated without human creativity and thus not eligible for copyright protection under law.

Copyright !== TOS. And any use of a service tends to constitute agreement to the terms.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#64

Post by dannyw »

codingdave wrote: post #62
huggingmouth wrote: post #54

Earlier quoted context omitted.

This is wrong. Terms of service only apply if you agree to them. They do do not magically apply if you download the weights from someone else. Both the weights and the fine-tuned data were generated without human creativity and thus not eligible for copyright protection under law.

Copyright !== TOS. And any use of a service tends to constitute agreement to the terms.

But you aren’t using someone else’s service. You are using model weights of something that is explicitly uncopyrightable; like public domain.

It may be against XYZ’s TOS to scrape public domain content from their website, but if you obtained that content elsewhere, XYZ’s TOS doesn’t apply to you.

It’s a weird world with AI and copyright.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#65

Post by anaganisk »

elaus wrote: post #3
Al-Khwarizmi wrote: post #2

The UI looks great, but I only get "Something went wrong, please try again." after a while waiting for the response :(

Probably immediately hugged to death by the HN crowd...

Hug of death for 137 upvotes post, Posted early in EST?

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#66

Post by dmd »

esperent wrote: post #48
dmd wrote: post #44

Me: How can I convince my 5 year old to put her shoes on in a timely fashion each morning? Answer in table form. LLaMA: 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. 1. 2. 3. 4. 5. (gpt: https://cloud.typingmind.com/share/0841633b-5150-4f7c-a370-0... )

> Answer in table form. What does this mean?

Look at my gpt link.

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#68

Post by vlovich123 »

dannyw wrote: post #64
codingdave wrote: post #62

Earlier quoted context omitted.

Copyright !== TOS. And any use of a service tends to constitute agreement to the terms.

But you aren’t using someone else’s service. You are using model weights of something that is explicitly uncopyrightable; like public domain. It may be against XYZ’s TOS to scrape public domain content from their website, but if you obtained that content elsewhere, XYZ’s TOS doesn’t apply to you. It’s a weird world with AI and copyright.

Dollars to donuts any judge would rule that the 7 billion weights constitute a copyright violation because you didn’t spontaneously come up with those weights on your own.

Heck, this “it’s just a number” logic didn’t pan out for DeCSS [1] and that was a fixed small encryption key where it wasn’t as clear cut. So why would things suddenly shift here where there’s a very clear line that this is a derived work from FB’s model? The only thing that would help you I think is if FB decided to not enforce their copyright but I don’t think they’ve made clarifying statements either way. If I recall correctly, unlike trademarks, copyrights don’t have requirements that you must go after all instances of infringement to maintain your rights so FB can be totally selective about which instances they choose to prosecute.

[1] https://en.m.wikipedia.org/wiki/Illegal_number

Re: Show HN: ChatLLaMA – A ChatGPT style chatbot for Facebook's LLaMA

#70

Post by weberer »

eob wrote: post #61
huggingmouth wrote: post #54

Earlier quoted context omitted.

This is wrong. Terms of service only apply if you agree to them. They do do not magically apply if you download the weights from someone else. Both the weights and the fine-tuned data were generated without human creativity and thus not eligible for copyright protection under law.

Has a court decided that yet? That’s huge if true. Someone once told me a curation was impossible to copyright for similar reasons.

It may apply to this case. The court ruled that Netscape's terms were not valid because end users did not give unambiguous consent before downloading. It may or may not apply here since people are downloading a leaked copy of Facebook's weights through a torrent.

https://cyber.harvard.edu/stjohns/Specht_v_Netscape.pdf

Post reply on HN