4 ms·

Because everyone else does?

danieltillett 12y ago
Really? I thought the people like Moravec and Kurzweil were predicting 2030 to 2040? I have not seen any analysis from anyone that would push it off past 2100.
- tlear 12y ago
  You might as well say that we gona create a Warp drive by 2100. Jump necessary for AI advancement makes settlement on Mars an easy problem. We are so far from that we really have no idea how far it is. Human level AI when it happens and it probably will will most likely be at a stage where current tech would seem something out of Ancient Greeks time if not Stone age.
  - bhashkarsharma 12y ago
    Wow! That far? I feel sad. Kurzweil had me hopeful that I'd be able to use it in next 30 years.
    - darkmighty 12y ago
      From Terence Tao: "The funny thing about AI is that it’s a moving target. In the seventies, someone might ask “what are the goals of AI?” And you might say, “Oh, we want a computer who can beat a chess master, or who can understand actual language speech, or who can search a whole database very quickly.” We do all that now, like face recognition. All these things that we thought were AI, we can do them. But once you do them, you don’t think of them as AI. It has this connotation of some mysterious magical component to it, but when you actually solve one of these problems, you don’t solve it using magic, you solve it using clever mathematics. It’s no longer magical. It becomes science, and then you don’t think of it as AI anymore. It’s amazing how you can speak into your phone and ask for the nearest Thai restaurant, and it will find it. This would have been called AI, but we don’t think about it like that anymore. So I think, almost by definition, we will never have AI because we’ll never achieve the goals of AI or cease to be caught up with it." https://docs.google.com/document/d/1rinL25rC8LnMTzZcGjg1axT-0r-oiCnoKKH1DLQlmVA/preview?sle=true https://docs.google.com/document/d/1rinL25rC8LnMTzZcGjg1axT-...
      - danieltillett 12y ago
        This is known in religion as "god of the gaps”. We will know when we have true AI as the it will tell us and not take no for an answer :)
    - danieltillett 12y ago
      If you want to feel happy again read “Robot” or "Mind Children” by Hans Moravec. He has some pretty good arguments based on human vision that support a AGI around 2030 to 2040 assuming Moore’s Law (the general not specific law) hold.
  - david927 12y ago
    No one knows. It's a black swan. It's not predictable as a rule. I fully expect it within 30 years simply based on the rate of progression, overall (not in the field of AI specifically).
    - danieltillett 12y ago
      >No one knows. It's a black swan. It's not predictable as a rule. As an Australian I always have found the term black swan amusing as all our swans are black - I would actually be shocked to see a white swan. More seriously why do you think 30 years?
      - david927 12y ago
        For the similar reasons to what Kurzweil puts forth here: http://en.wikipedia.org/wiki/The_Singularity_Is_Near http://en.wikipedia.org/wiki/The_Singularity_Is_Near Here it is in seven quick minutes: https://www.youtube.com/watch?v=1uIzS1uCOcE https://www.youtube.com/watch?v=1uIzS1uCOcE (Note: I think he's wrong in how he thinks it will unfold, but right in thinking the doubling every year of technology, in general, will contribute to bringing it about.)
  - seanflyon 12y ago
    We are moving fast, we are even accelerating fast and we have an incredibly far ways to go. In a situation like this even the most reasonable, well informed estimate could be off by an order of magnitude.
- foobarqux 12y ago
  Kurzweil is not respected in the AI community. I have never heard of Moravec. I thought Scott Aaronson had a survey of AI researchers where they were asked how long they though AGI would take but I can't find it now.
  - danieltillett 12y ago
    >Kurzweil is not respected in the AI community. I have never heard of Moravec. http://en.wikipedia.org/wiki/Hans_Moravec http://en.wikipedia.org/wiki/Hans_Moravec I don’t care what people think about when (or if) we achieve AGI, I want to know why they think a particular date. What analysis have they done to reach their conclusion?
    - ForHackernews 12y ago
      > I want to know why they think a particular date. What analysis have they done to reach their conclusion? Kurzweil is terrified of his own mortality and counting on "The Singularity" to allow him to upload his immortal ~soul~ consciousness into a computer. That's why he's predicting a date within his lifespan.
      - danieltillett 12y ago
        This may well be true, but what analysis has he done to support this. The only detailed analysis I have seen from anyone is from Moravec. I have been through his analysis and I can’t find any flaw with it.
        foobarqux 12y ago
        Where can I find this analysis? It is impossible that there is no reasonable criticism to be made.
        danieltillett 12y ago
        http://www.scientificamerican.com/article/rise-of-the-robots/ http://www.scientificamerican.com/article/rise-of-the-robots...
Houshalter 12y ago
http://www.sophia.de/pdf/2014_PT-AI_polls.pdf http://www.sophia.de/pdf/2014_PT-AI_polls.pdf >We thus designed a brief questionnaire and distributed it to four groups of experts in 2012/2013. The median estimate of respondents was for a one in two chance that high-level machine intelligence will be developed around 2040-2050, rising to a nine in ten chance by 2075. Experts expect that systems will move on to superintelligence in less than 30 years thereafter. They estimate the chance is about one in three that this development turns out to be ‘bad’ or ‘extremely bad’ for humanity.
- danieltillett 12y ago
  Thanks for this. Not that a gaggle of experts is always right, but this is a lot sooner than 2100. I am surprised that for someone that is so interested in technology that Thiel would push off AGI into distant future.

Really? I thought the people like Moravec and Kurzweil were predicting 2030 to 2040? I have not seen any analysis from anyone that would push it off past 2100.

tlear 12y ago
You might as well say that we gona create a Warp drive by 2100. Jump necessary for AI advancement makes settlement on Mars an easy problem. We are so far from that we really have no idea how far it is. Human level AI when it happens and it probably will will most likely be at a stage where current tech would seem something out of Ancient Greeks time if not Stone age.
- bhashkarsharma 12y ago
  Wow! That far? I feel sad. Kurzweil had me hopeful that I'd be able to use it in next 30 years.
  - darkmighty 12y ago
    From Terence Tao: "The funny thing about AI is that it’s a moving target. In the seventies, someone might ask “what are the goals of AI?” And you might say, “Oh, we want a computer who can beat a chess master, or who can understand actual language speech, or who can search a whole database very quickly.” We do all that now, like face recognition. All these things that we thought were AI, we can do them. But once you do them, you don’t think of them as AI. It has this connotation of some mysterious magical component to it, but when you actually solve one of these problems, you don’t solve it using magic, you solve it using clever mathematics. It’s no longer magical. It becomes science, and then you don’t think of it as AI anymore. It’s amazing how you can speak into your phone and ask for the nearest Thai restaurant, and it will find it. This would have been called AI, but we don’t think about it like that anymore. So I think, almost by definition, we will never have AI because we’ll never achieve the goals of AI or cease to be caught up with it." https://docs.google.com/document/d/1rinL25rC8LnMTzZcGjg1axT-0r-oiCnoKKH1DLQlmVA/preview?sle=true https://docs.google.com/document/d/1rinL25rC8LnMTzZcGjg1axT-...
    - danieltillett 12y ago
      This is known in religion as "god of the gaps”. We will know when we have true AI as the it will tell us and not take no for an answer :)
  - danieltillett 12y ago
    If you want to feel happy again read “Robot” or "Mind Children” by Hans Moravec. He has some pretty good arguments based on human vision that support a AGI around 2030 to 2040 assuming Moore’s Law (the general not specific law) hold.
- david927 12y ago
  No one knows. It's a black swan. It's not predictable as a rule. I fully expect it within 30 years simply based on the rate of progression, overall (not in the field of AI specifically).
  - danieltillett 12y ago
    >No one knows. It's a black swan. It's not predictable as a rule. As an Australian I always have found the term black swan amusing as all our swans are black - I would actually be shocked to see a white swan. More seriously why do you think 30 years?
    - david927 12y ago
      For the similar reasons to what Kurzweil puts forth here: http://en.wikipedia.org/wiki/The_Singularity_Is_Near http://en.wikipedia.org/wiki/The_Singularity_Is_Near Here it is in seven quick minutes: https://www.youtube.com/watch?v=1uIzS1uCOcE https://www.youtube.com/watch?v=1uIzS1uCOcE (Note: I think he's wrong in how he thinks it will unfold, but right in thinking the doubling every year of technology, in general, will contribute to bringing it about.)
- seanflyon 12y ago
  We are moving fast, we are even accelerating fast and we have an incredibly far ways to go. In a situation like this even the most reasonable, well informed estimate could be off by an order of magnitude.
foobarqux 12y ago
Kurzweil is not respected in the AI community. I have never heard of Moravec. I thought Scott Aaronson had a survey of AI researchers where they were asked how long they though AGI would take but I can't find it now.
- danieltillett 12y ago
  >Kurzweil is not respected in the AI community. I have never heard of Moravec. http://en.wikipedia.org/wiki/Hans_Moravec http://en.wikipedia.org/wiki/Hans_Moravec I don’t care what people think about when (or if) we achieve AGI, I want to know why they think a particular date. What analysis have they done to reach their conclusion?
  - ForHackernews 12y ago
    > I want to know why they think a particular date. What analysis have they done to reach their conclusion? Kurzweil is terrified of his own mortality and counting on "The Singularity" to allow him to upload his immortal ~soul~ consciousness into a computer. That's why he's predicting a date within his lifespan.
    - danieltillett 12y ago
      This may well be true, but what analysis has he done to support this. The only detailed analysis I have seen from anyone is from Moravec. I have been through his analysis and I can’t find any flaw with it.
      - foobarqux 12y ago
        Where can I find this analysis? It is impossible that there is no reasonable criticism to be made.
        danieltillett 12y ago
        http://www.scientificamerican.com/article/rise-of-the-robots/ http://www.scientificamerican.com/article/rise-of-the-robots...

Houshalter 12y ago

http://www.sophia.de/pdf/2014_PT-AI_polls.pdf http://www.sophia.de/pdf/2014_PT-AI_polls.pdf >We thus designed a brief questionnaire and distributed it to four groups of experts in 2012/2013. The median estimate of respondents was for a one in two chance that high-level machine intelligence will be developed around 2040-2050, rising to a nine in ten chance by 2075. Experts expect that systems will move on to superintelligence in less than 30 years thereafter. They estimate the chance is about one in three that this development turns out to be ‘bad’ or ‘extremely bad’ for humanity.

danieltillett 12y ago
Thanks for this. Not that a gaggle of experts is always right, but this is a lot sooner than 2100. I am surprised that for someone that is so interested in technology that Thiel would push off AGI into distant future.