4 ms·

I don't think we can try to tell companies like Facebook that they're being irresponsible and unethical and then ask that every tech company releases tech that

• nharada 6y ago
  I actually agree with this, and I thought that OpenAI's decision to not open source GPT-2 was smart and measured at the time. I didn't fully believe the commentary that it was profit motivated. But I believe that single actions should be viewed in the context of patterns of behavior. Truthfully, deciding not to open source GPT-2 can be both profit motivated and safety oriented. I had thought it was mostly the latter, but in the context of the company's choices it looks like more the former than I thought.
  • phkahler 6y ago
    Yep. If AI is a threat to anyone it's the size and reach of those who control it that makes it potentially dangerous. If that's a concern you don't do this. It's all about the Benjamins.
    • slg 6y ago
      >If AI is a threat to anyone it's the size and reach of those who control it that makes it potentially dangerous. The important part is not the size and reach of the controlling entity. It is the motives of that entity. Microsoft is motivated by profit, but they are also motivated to continue to legally operate in almost every country in the world. They aren't going to use this technology in blatantly illegal ways and probably won't use it in actively harmful ways. Open sourcing this tech would allow it to fall into much more nefarious hands.