Skip to content


The singularity actually is near

I finished Ray Kurzweil’s book on the Singularity last week and finally got around to reviewing it. It turned out to be extraordinarily long. Go read it!

Posted in Books.


5 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Bryan says

    Ok, so it would be awesome if we all became cyborgs. But… this guy seems a bit overly optimistic about the possibility of this new epoch. As far as I can tell, we’re just as far away from understanding conscious intelligence as we were 100 years ago — and no closer to augmenting it at all.

  2. Zach Musgrave says

    We’re definitely way closer to understanding intelligence than we were 100 years ago. If you accept that our knowledge is increasing exponentially — and Kurzweil makes a compelling case that it is — then we’re even closer than a simple linear projection would predict. 100 years ago, we barely understood that nerve impulses were electric. Today, we can visualize electrical activity in the brain with 1 mm and 1s resolution. We can control mice via remote control (and implants). We can reproduce entire brain regions as a black-box circuit; cochlear implants were the first commercial application. So while it’s true that we have a long way to go, our progress is truly astounding. I see no reason why it shouldn’t continue, and that path must lead to an understanding of the physical mechanism of consciousness. There is one, you sucka!

    I’ve been told that the Age of Spiritual Machines is his good book, so you might want to pick that one up instead if you were planning on reading his arguments in more detail.

  3. Bryan says

    Look, these are very good points. I really am with you that we now deeply understand the relationship between brain activity and motor reaction, and we didn’t 100 years ago. But here’s all that amounts to: we’re becoming great puppeteers. And augmenting a puppet isn’t augmenting intelligence.

    Your first point, on visualizing brain-activity: true, fMRIs are this amazing new development. But their interpretation — and especially their relation to actual mental states — is completely up in the air. The visual image they provide is actually very misleading. Those images are the result of enumerable noise-reduction processes and colorful theorizing. And even if we do trust them, there is little evidence that just because “Region X” lights up when we do Y, that Region X is also the *cause* of us doing Y.

    As for a physical mechanism behind consciousness… that’s really interesting. The problem with the whole research project seems to me to be that no one knows what “consciousness” (or “intelligence,” for that matter) even means. There are lots of pithy metaphors, but has any precise connection every been made? Maybe it’s even the kind of thing that *can’t* be defined precisely. Or maybe it’s just puppeteering (although that seems unsatisfying), in which case I suppose I’d agree with you that we are making progress. At any rate, I’m holding back any hope for a science of consciousness until this basic foundational question is cleared up.

  4. Zach Musgrave says

    Personally, my bet is that we succeed in reverse-engineering the brain’s major functions, including “intelligence” and “consciousness,” before we truly understand the mechanism behind those functions. We’ll have silicon brains that pass the Turing test, but we won’t understand why.

    In any case, if Kurzweil is even remotely on base with his time line predictions, this question will be settled by the early 2030s. Care to make a bet? A bottle of (non-nanobot-produced) Brandy that we’ve passed the Turing test by 2030?

    • Bryan says

      Hmm, that’s tempting. Notably, passing a Turing test may have little to do with consciousness. But it is a precise enough definition that we might actually settle a bet. We suck at the Turing test now. Not much improvement in the last 10 years, either. Ok, you’re on.

      However, a 21-year bottle of scotch seems more appropriate, given the circumstances…



Some HTML is OK

or, reply to this post via trackback.