We've trained an unsupervised language model that can generate coherent paragraphs and perform rudimentary reading comprehension, machine translation, question answering, and summarization — all without task-specific training: https://blog.openai.com/better-language-models/ …pic.twitter.com/360bGgoea3
-
What you are doing is opposite of open. It is unfortunate that you hype up +propagate fear + thwart reproducibility+scientific endeavor. There is active research from other groups in unsupervised language models. You hype it up like it has never been done before.
@jackclarkSF
I summarize @OpenAI controversy and my stance here. @ilyasut @zacharylipton @WWRob @Miles_Brundage @tsimonite @jjvincent @glenweyl @rasbt @beenwrekt @smerity @samcharrington @ChiefScientist @egrefen https://anima-ai.org/2019/02/18/an-open-and-shut-case-on-openai/ …
-
-
I think
@OpenAI did the right thing It creates friction between people that want to use it Otherwise, every single teenager could use the model directly to troll I know some people on 4chan that would be estatic of this -
This Tweet is unavailable.
- Show replies
New conversation -
-
-
The model is not revolutionary. The main innovation is training data-collection. Should be easy to reproduce. The point of holding back is not to be closed, but to allow time to prepare for a potential threat. Similar to bug reports for open source software. Seems reasonable.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thanks
@AnimaAnandkumar , as always you express clearly what it's not yet clear in my mind! P.S. this episode is one more reason why something like the@PeerCommunityIn project is important ;-)Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
While I agree it is good to discuss. I believe they have simply perpetuated and deepened the media and hence public bias that all AI is bad. This was badly handled by OpenAI and made worse by their invitation for government oversight over self regulation
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I think
@OpenAI/@jackclarkSF agree that "increasing reproducibility and accountability" is important - but the argument here seems to say you cannot balance this with any other concerns, and so the only acceptable path is full transparency. Or am I misunderstanding this? -
Regarding your call for
@OpenAI to "engage with experts on risk management to study the impacts of AI," I know that they agree, because this is a large and ongoing project. (I'm peripherally involved, but OpenAI has been closely involved in doing exactly this since its founding.)
End of conversation
New conversation -
-
-
Has it crossed your mind that given too much controversy and outrage, and some research labs might just decide not to publish at all and keep progress internal? If you're big enough, you could choose to do that
-
I’d personally not open a dialogue with a professor with “have it crossed your mind”, as all kinds of ideas cross their minds all the time... just sayin’:)
- Show replies
New conversation -
-
-
Your website says you are a professor in "Arfiticial intelligence"
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.