News for hackers. Stuff that matters.

A comment on OpenGPT-2: We Replicated GPT-2 Because You Can Too. Parent comment

Re:OpenGPT-2: We Replicated GPT-2 Because You Can Too by anchpop on Friday August 23, 2019 @09:50AM (#20777569)
Before they release it, they need to prove to themselves it's safe (I'm speaking morally). I'm not sure why so many on HN seem to think it's the other way around, that you have to prove it's dangerous before withholding it. $500k for reproduction is peanuts for a nation state but quite a bit for many other groups who might be interested.

Even if it is entirely safe, it's still good to withhold it because it helps create a culture where people think about the safety of their projects before releasing them

[ Parent | on HN ]
Re:OpenGPT-2: We Replicated GPT-2 Because You Can Too by Bartweiss on Friday August 23, 2019 @11:00AM (#20778327)
There are two very different questions here: is it good to withhold, and is that sensible for OpenAI to withhold? I think a lot of people on HN view restricting GPT as incoherent by OpenAI, where they might not for another group.

OpenAI's raison d'etre was democratizing access to AI tools, and preventing them from being abused by concentrated powers like governments. If replicating GPT-2 is trivial for state actors but prohibitive for hobbyists and other private citizens, it's creating the same issue Musk described OpenAI as setting out to oppose. Even the general idea of treating AI tools as hazardous-by-default goes a long way to validating the project's original critics.

[ Parent | on HN ]
Re:OpenGPT-2: We Replicated GPT-2 Because You Can Too by buboard on Friday August 23, 2019 @11:14AM (#20778472)
ITs really difficult to impossible to prove it s eithrt safe or not. We re still in the early stage of ML research so this kind of fud is a brake to research. Similar to how genetic research has been held back by regulators
[ Parent | on HN ]
Re:OpenGPT-2: We Replicated GPT-2 Because You Can Too by jgalt212 on Friday August 23, 2019 @10:39AM (#20778100)
> Before they release it, they need to prove to themselves it's safe (I'm speaking morally)

I don't think it was a moral question. I think the lawyers got involved and were uncomfortable with potential litigation exposure.

[ Parent | on HN ]