RSS

Episode 44: Daniel Dewey – Thinking Carefully About Artificial Intelligence

26 Mar

In this episode, Gabe and Dan have the opportunity to chat with Daniel Dewey about artificial intelligence (AI) and intelligence explosion. Daniel is a researcher at the Future of Humanity Institute, researching artificial intelligence, reinforcement learning, and how machines could have values.

Check out Daniel’s TEDx talk: The long-term future of AI (and what we can do about it)

Listen to this episode

Some links to what we talked about:

The book we mentioned in the episode:

Links to other people whose cool research came up in the discussion:

Thanks for listening!

About these ads
 
 

Tags: , ,

8 responses to “Episode 44: Daniel Dewey – Thinking Carefully About Artificial Intelligence

  1. Evan

    March 29, 2014 at 1:12 am

    Is this supposed to be a real topic? Plz tell me youre not taking seriously the evil robots killing all humans story. Yes it makes for fun scifi, but little else.

     
    • nontheology

      March 29, 2014 at 3:48 am

      Haha, yeah it’s a real topic, although I don’t claim everyone is going to find it interesting. I think the thought experiments and reasoning are powerful, which we didn’t get too deep into in the discussion.. Why do you think it’s limited to sci-fi?

       
      • Evan

        March 30, 2014 at 2:29 am

        It’s not that it isnt interesting, I was attracted to the same themes in fiction when I was young, and I still think they make for a great read. But I take it about as seriously as I do Star Trek. I don’t remember what episode it was, but Gabe had said something about Craig’s cosmological argument that if you can sit around in your armchair and figure out necessary facts of what created the universe, then your doing science wrong… Why do you say that about Craig, but then Sit down and talk about how robots are going to take over the world? The first completely abandons empiricism, and it seems like the second does the same.
        The guest makes my case during the show too when he says that people in this field have a bad track record of making accurate predictions. If you can’t make accurate predictions, and you’re aware of that, you shouldnt be expecting people to take you seriously when you tell them that they should be worried about a threat that doesn’t exist yet.

         
      • Evan

        March 30, 2014 at 2:33 am

        like saying “I’m bad at making predictions, and I predict this bad stuff is gonna happen, so you should buy my book to do something about it” no, I’m not going to bother with that.

         
      • nontheology

        March 30, 2014 at 3:05 am

        Evan:

        Yeah, your comparison to my criticism of the Kalam argument is interesting. Sharp ears. To some extent, I agree that the worry about AI isn’t strictly founded on evidence of AI harming humans, but I also think that when we’re talking about something that could potentially end humanity, we shouldn’t be waiting around for the evidence that it will actually do so. I feel the same about climate change phenomena: we shouldn’t wait around for humans to create a runaway greenhouse affect in the atmosphere before we worry about out greenhouse gas output. If the claim is that “X might destroy humanity”, and we care about humanity, we shouldn’t wait to see whether X does in fact destroy humanity, especially if we have strong arguments to suggest that such a situation is plausible given what was know about X. And I think that’s the case with AI and intelligence explosion.

        As for your point about bad predictions: I was actually glad that Daniel made the comment about the track record of predictions, because non-trivial predictions are hard to make and a lot of people don’t seem to be aware of that. But I took that to be more about dates that something will happen, than that predictions themselves are inaccurate (though I’m sure there’s plenty of that as well). It’s not that the field can’t make accurate predictions, just that the timing isn’t spot on. I don’t know, I think you’re taking that to mean something other than what was intended.

         
  2. xxicenturyboy

    July 5, 2014 at 3:19 pm

    So when are you putting out a new pod! Your listener wants to know…

     
    • nontheology

      July 5, 2014 at 3:44 pm

      Argh. I have an episode recorded and ready to post, but my internet is down! You’ll see it up when I get ones and zeros flowing back into my house again! Stupid hurricanes…

       

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
Follow

Get every new post delivered to your Inbox.

Join 49 other followers

%d bloggers like this: