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

OpenAI releasing new open model in coming months, seeks community feedback

openai.com

Post reply on HN
Print view View original

41–50 of 82 posts

Re: OpenAI releasing new open model in coming months, seeks community feedback

#41

Post by impossiblefork »

vessenes wrote: post #34

For those wondering how to answer "what do you want to see from an open model" I put this in: an open weights end to end multimodal model, a model large enough to act as a functional teacher along with a range of nicely distilled smaller sizes, code repo to make training / finetuning easy. As I write this, I'd also like to request a set of tool calling LLMs in various sizes. Feels to me like a small fast local tool c…

I think that when learning privacy and being able to run things oneself is less of a problem.

Keeping ones work confidential becomes interesting once one starts doing new things, so while it would be fun, I don't think that's the most useful type of open model. I think code models are probably the most useful.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#42

Post by fooker »

bhouston wrote: post #12

Q: How open? Can it be used commercially? Is the training protocol going to be open? Or just the weights released like Llama models?

It's a kind of an open secret that there's no 'training' protocol for these state of the art models.

Researchers behave like alchemists when training these models, and the actions are not really reproducible.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#43

Post by Spivak »

BoredPositron wrote: post #36

Don’t kid yourself about the intentions. It’s only so enterprise customers can deploy a shitty chatbot on premise for their "secret data". The models will be free for commercial use until 1 mil turnover or something like that.

I mean for large enough buyers that was already an option, they wouldn't need to release a new model just for that.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#44

Post by lawlessone »

vntok wrote: post #29
lawlessone wrote: post #24

Earlier quoted context omitted.

>they need to make a selection removing trolls, luddites, etc Do the luddites opinions not count?

To the tech sector's asking for input to help shape their vision? I would guess they don't.

I just don't think a word predictor is the future of AI.

It seems like a sloppy way to simulate intelligence.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#45

Post by ewalk153 »

Liwink wrote: post #33

What's the major use case of open source models? The stable diffusion community seems pretty active. A lot of fine tuning to generate NSFW. What about LLM?

When coupled with smaller parameters sizes, it enables BYOD, different cost scaling, and local inference.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#46

Post by pests »

kopecs wrote: post #39
wewtyflakes wrote: post #37

Earlier quoted context omitted.

If you have a business selling hamburgers, does it make sense to ask vegetarians what should be on it?

I would think it makes some amount of sense if you think they're vegetarian for some moral reason and you think you could court them to become customers.

But why target a market that doesn't want you? Do you really think a burger place can some how override a person's morals to not eat meet? I realize were stretching the analogy here, but what is the point? Maybe if you already saturated the existing market and trying to grow. But then why saddle that market with the wants/needs of the polar opposite?

Re: OpenAI releasing new open model in coming months, seeks community feedback

#47

Post by 9rx »

wewtyflakes wrote: post #37
lawlessone wrote: post #24

Earlier quoted context omitted.

>they need to make a selection removing trolls, luddites, etc Do the luddites opinions not count?

If you have a business selling hamburgers, does it make sense to ask vegetarians what should be on it?

Given that they come in good faith (trolling is already excepted), I would ask. Burger toppings are predominately plant-based, so perfectly in their wheelhouse. In fact, I'd expect better suggestions from them than the average burger eater.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#48

Post by rockinghigh »

Spivak wrote: post #43
BoredPositron wrote: post #36

Don’t kid yourself about the intentions. It’s only so enterprise customers can deploy a shitty chatbot on premise for their "secret data". The models will be free for commercial use until 1 mil turnover or something like that.

I mean for large enough buyers that was already an option, they wouldn't need to release a new model just for that.

I don't believe OpenAI provides on-premise serving.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#49

Post by stale2002 »

lawlessone wrote: post #24
vntok wrote: post #16

Earlier quoted context omitted.

Well obviously they will have thousands of applications from which they need to make a selection removing trolls, luddites, etc. How would you do it without asking for some applicants' details?

>they need to make a selection removing trolls, luddites, etc Do the luddites opinions not count?

If their plan is to sabotage things, then no their opinions should not count.

Re: OpenAI releasing new open model in coming months, seeks community feedback

#50

Post by daemonologist »

Liwink wrote: post #33

What's the major use case of open source models? The stable diffusion community seems pretty active. A lot of fine tuning to generate NSFW. What about LLM?

It's a huge compliance hurdle at my company to add any new vendor, so we almost exclusively use open models that we can run on our own hardware (or our rented cloud instances). Even just getting Bedrock enabled in one of our existing AWS accounts has been in the works for months.
Post reply on HN