Live data from Hacker News

ChatGPT Enterprise

openai.com

461–470 of 532 posts

Re: ChatGPT Enterprise

#461

Earlier quoted context omitted.

Totally agree. retrieval augmented generation is still the preferred way to give the LLM more knowledge. Fine-tuning is mostly useful for adapting the base model for another task. I wrote about this in a recent blog post: https://vectara.com/fine-tuning-vs-grounded-generation/ . Anyone knows how this new capability works in terms of where the model inference be done? Would it still be at the OpenAI side or is this go…

In your opinion, is it an either or scenario? Or would fine-tuning on docs + RAG be even more powerful?

I've been wondering this myself lately.

After using RAG with pgvector for the last few months with temperature 0, it's been pretty great with very little hallucination.

The small context window is the limiting factor.

In principle, I don't see the difference between a bunch of fine-tuned prompts along the lines of "here is another context section: ", which is the same as what it looks like in a RAG prompt anyway.

Maybe the distinction of whether it is for "tone" or "context" is based on the role of the given prompts and not restricted by the fine-tuning process itself?

In theory, fine-tuning it on ~100k tokens like that would allow for better inference, even with the RAG prompt that includes a few sections from the same corpus. It would prevent issues where the vector search results are too thin despite their high similarity. E.g. picking out one or two sections of a book which is actually really long.

For example, I've seen some folks use arbitrary chunking of tokens in batches of 1k or so as an easy config for implementation, but that totally breaks the semantic meaning of longer paragraphs, and those paragraphs might not come back grouped together from the vector search. My approach there has been manual curation of sections allowing variations from 50 to 3k tokens to get the chunks to be more natural. It has worked well but I could still see having the whole corpus fine-tuned as extra insurance against losing context.

Re: ChatGPT Enterprise

#462

Well the message in this video certainly did not age well: https://www.youtube.com/watch?v=smHw9kEwcgM TLDR: This might have just killed a LOT of startups

Any startup that is using ChatGPT under the hood is just doing market research for OpenAI for free. The same happened when people started experimented with GPT3 for code completion, right before being replaced by Copilot. If you want to build an AI start-up and need a LLM, you must use Llama or another model than you can control and host yourself, anything else is basically suicide.

>Any startup that is using ChatGPT under the hood is just doing market research for OpenAI for free

It's not free if you have paying clients.

> If you want to build an AI start-up and need a LLM, you must use Llama or another model than you can control and host yourself, anything else is basically suicide.

You're still doing market research for OpenAI. Just because you aren't using their model doesn't mean they can't copy your UX. Prompts are not viable trade secrets after all.

Re: ChatGPT Enterprise

#463
post #401

Earlier quoted context omitted.

> Are you claiming this because they used copyrighted material as training data? If so, I think you're starting from the wrong point. All open source license comes under copyright law. It means if they violate the OSS license, the license is void and the tech/material becomes copyright protected. So yes, it would mean that it is trained on copyrighted material. > Additionally, I don't think many open source licenses…

If they violate their own terms they will lose customers and money. If they violate OSS nothing happens.

If they violate their own terms, nothing happens until someone knows about it and is inclined to act against them.

Re: ChatGPT Enterprise

#464

Earlier quoted context omitted.

> So why do we care from where LLMs learn? Because humans aren't computers and the similarities between the two, other than the overuse of the word "learning" in the computer's case, are nonexistant?

Are you really asserting that these models aren't learning? What definition of learning are you using?

Don't know if they are, and don't really care either and I don't care to anthropomorphize circuitry to the extent that AI proponents tend to, especially.

Humans and Computers are 2 wholly separate entities, and there's 0 reason for us to conflate the two. I don't care if another human looks at my code and straight up copies/pastes it, I care very much if an entity backed by a megacorp like Micro$oft does the same, en-masse, and sells it for profit, however.

Re: ChatGPT Enterprise

#465

Earlier quoted context omitted.

Their decision makes sense, in a weird way. A lot of value in some SaaS apps is in the initial investment it took to build it, not in the cost to host a customer's assets. If the runtime costs of a new customer are negligible, would you rather have 0K or 20K?

Of course, I'd rather have 20K per customer But an initial quote of 300K would likely lead to many instant rejections rather than engaging in negotiation, right? That's why I say it feels like a stupid practice, even though it could pay off really well if some company accepts outright (With the caveat that I've never been near this kind of business deal, so I'm just going off of common sense)

Sure, but some people will accept the 300K so it could be worth it even if you scare off a majority of your potential customers.

If dodgy pricing/sales tactics didn't work then Oracle would be bankrupt instead of a 300 billion dollar company.

Re: ChatGPT Enterprise

#466

Earlier quoted context omitted.

> So why do we care from where LLMs learn? Because humans aren't computers and the similarities between the two, other than the overuse of the word "learning" in the computer's case, are nonexistant?

Are you really asserting that these models aren't learning? What definition of learning are you using?

Do humans really read terabytes of C code to learn C?

Humans look at a few examples and extrapolate…

Re: ChatGPT Enterprise

#467
If your company illegally treads on my IP, I don't care if your employees used LLM's or not; I will sue.

By the way, I don't have to win a lawsuit to get some justice; I'll make discovery hurt.

So if OpenAI stole my IP and used it for training (which they probably did, illegally IMO), I guess you're taking that risk if you let your employees use their LLM's.

Re: ChatGPT Enterprise

#468

Earlier quoted context omitted.

My company (Fortune 500 with 80,000 full time employees) has a policy that forbids the use of any AI or LLM tool. The big concern listed in the policy is that we may inadvertently use someone else’s IP from training data. So, our data going into the tool is one concern, but the other is our using something we are not authorized to use because the tool has it already in its data. How do you prove that that could never…

> How do you prove that that could never occur? Realistically you can prove that just as well as you can prove that employees aren't using ChatGPT via their cellphones. There are also organizations that forbid the use of Stack overflow. As long as employees don't feel like you're holding back their career and skills by prohibiting them from using modern tools, and keep working there, hey. As long as you pay them enou…

You can't code without access to stackoverflow?

Official documentation is still available…

Re: ChatGPT Enterprise

#469

Earlier quoted context omitted.

I don't think using ChatGPT is similar to searching for answers on S.O. Maybe if you were asking people on S.O. to write your code for you, or plugging in exact snippets. The point here is that letting ChatGPT write code directly into your repo is effectively plagiarism and may violate any number of licenses you don't even realize you're breaking, whereas just looking at how other people did something , understanding…

I've been experiencing carpal tunnel on and off for a couple of weeks now. I can tell you that reading through some code generated by "insert llm x" is substantially less painful than writing all of it by my own hand. Especially if you start understanding how to refine your prompts to the point where you use a single thread for project management and use that thread to generate prompts for other threads. Not all valu…

Take sick leave

Re: ChatGPT Enterprise

#470

Earlier quoted context omitted.

Using chatgpt to code is not a skill. It’s a crutch. Any employees that feel held back by not being able to access it aren’t great in the first place.

I remember people saying literally the exact same thing almost word for word about Google almost a quarter century ago.

You mean the search engine that NEVER EVER gives me the documentation I'm locking for, but always goes for a vaguely related blog with hundreds of ads?
Post reply on HN