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

OpenAI API

beta.openai.com

Post reply on HN
Print view View original

71–80 of 165 posts

Re: OpenAI API

#71

Post by zitterbewegung »

minimaxir wrote: post #24
zitterbewegung wrote: post #17

Looks like OpenAI is going head to head with huggingface. This makes a lot of sense and it seems they are telegraphing to monetize what they have been doing. It also seems like this is why they don't release their models in a timely manner.

The notable difference is that the base Huggingface library is open source, so you could in theory build something similar or more custom to the OpenAI API internally (which then falls into the typical cost/benefit analysis of doing so).

So its like Github vs Gitlab which makes more sense. I can see huggingface have a hosted version because now you can share your models on their platform.

Re: OpenAI API

#72

Post by minimaxir »

swyx wrote: post #67
eggsnbacon1 wrote: post #47

OpenAI started as a non-profit, went for-profit. Still owned by the big players.... Something isn't right. Is OpenAI just a submarine so the tech giants can do unethical research without taking blame??? Its textbook misdirection, nonprofit and "Open" in the name, hero-esque mission statement. How do you make the mental leap from "we're non-profit and we won't release things too dangerous" to "JK we're for-profit and…

wow you just made the connection for me. GPT2 was too dangerous to release, and now GPT3 is so much better - is there no point at which things become too dangerous anymore? what was the conclusion on that one?

The blog post directly addresses this question: https://openai.com/blog/openai-api/

> What specifically will OpenAI do about misuse of the API, given what you’ve previously said about GPT-2?

> We will terminate API access for use-cases that cause physical or mental harm to people, including but not limited to harassment, intentional deception, radicalization, astroturfing, or spam; as we gain more experience operating the API in practice we expect to expand and refine these categories.

Re: OpenAI API

#73

Post by sailingparrot »

eggsnbacon1 wrote: post #47

OpenAI started as a non-profit, went for-profit. Still owned by the big players.... Something isn't right. Is OpenAI just a submarine so the tech giants can do unethical research without taking blame??? Its textbook misdirection, nonprofit and "Open" in the name, hero-esque mission statement. How do you make the mental leap from "we're non-profit and we won't release things too dangerous" to "JK we're for-profit and…

> so the tech giants can do unethical research

I know it's trendy (and partly justified) to look down on OpenAI, but can you actually give any basis for this claim?

What kind of research is OpenAI doing that all the other big AI players (Google/DeepMind, FB, Microsoft) aren't also invested in? And even if others are doing the same, what part of OpenAI's research do you consider unethical?

> It may just be a conspiracy theory

Yea, it very much looks like that to be honest.

Re: OpenAI API

#74

Post by zitterbewegung »

eggsnbacon1 wrote: post #47

OpenAI started as a non-profit, went for-profit. Still owned by the big players.... Something isn't right. Is OpenAI just a submarine so the tech giants can do unethical research without taking blame??? Its textbook misdirection, nonprofit and "Open" in the name, hero-esque mission statement. How do you make the mental leap from "we're non-profit and we won't release things too dangerous" to "JK we're for-profit and…

GPT-2 is hard to do "shady" things right now (speaking from experience)[1] but maybe GPT-3 might do better?

I could get poems to generate well. Tweets were a bit harder but I don't think we are at the point where you could just use a generative model to fool people that would be cheaper than actually hiring someone to write fake news. (Also shameless plug below)

[1] 1400 - TALK.8 - "A way to make fake tweets using GPT2" - Joshua Jay Herman https://thotcon.org/schedule.html

Re: OpenAI API

#75

Post by Yajirobe »

dna_polymerase wrote: post #56
chillee wrote: post #52

Earlier quoted context omitted.

Why is it "impossible"? Academic labs are non-profit, and they are also effectively trying "to compete" with DeepMind.

Have a look at this discussion and the article from earlier today [0]. Of course, a singular lab could compete with something DeepMind does, but not without massive amounts of money in their pockets. The state of the art has become pretty expensive, really fast. [0]: https://news.ycombinator.com/item?id=23486163

State of the art can be (and usually is) born in academic labs.

Re: OpenAI API

#76

Post by gdb »

say_it_as_it_is wrote: post #66

OpenAI started off wide-eyed and idealistic but it made the mistake of taking on investors for a non-profit mission. A non-profit requires sponsors, not investors. Investors have a fiduciary responsibility to maximize profits, not achieve social missions of open AI for all.

OpenAI LP, our "capped-profit" entity which has taken investment, has a fiduciary duty to the OpenAI Charter: https://openai.com/blog/openai-lp/

Re: OpenAI API

#77

Post by dmvaldman »

azinman2 wrote: post #63
dmvaldman wrote: post #46

AGI in text is < 3yrs away.

What breakthrough occurred?

Zero shot and few-shot learning in GPT-3 and lack of significant diminishing returns in scaling text models. Zero-shot learning is equivalent to saying "i'm just going to ask the model something that it was not trained to do"

Re: OpenAI API

#78

Post by azinman2 »

dmvaldman wrote: post #77
azinman2 wrote: post #63

Earlier quoted context omitted.

What breakthrough occurred?

Zero shot and few-shot learning in GPT-3 and lack of significant diminishing returns in scaling text models. Zero-shot learning is equivalent to saying "i'm just going to ask the model something that it was not trained to do"

And how do we get from zero shot to AGI? You're making gigantic leaps here.

Re: OpenAI API

#79

Post by dmvaldman »

Barrin92 wrote: post #62
dmvaldman wrote: post #46

AGI in text is < 3yrs away.

there's zero understanding in any of this. This is still just superficial text parsing essentially. Show me progress on Winograd schema and I'd be impressed. It hasn't got anything to do with AGI, this is application of ML to very traditional NLP problems.

i think you are assuming that what is happening under the hood is that a human-inputted sentence is being parsed into a grammar. it is not.

Re: OpenAI API

#80

Post by deep_etcetera »

simonkafan wrote: post #70
swyx wrote: post #67

Earlier quoted context omitted.

wow you just made the connection for me. GPT2 was too dangerous to release, and now GPT3 is so much better - is there no point at which things become too dangerous anymore? what was the conclusion on that one?

GPT2 being "too dangerous to release" was a marketing stunt from the very beginning.

Who are you quoting here?
Post reply on HN