5 ms·

Note that a very significant fraction do believe human level machine intelligence is likely within about 50 years, and a good few on closer timescales, though t

Note that a very significant fraction do believe human level machine intelligence is likely within about 50 years, and a good few on closer timescales, though the precise numbers are endlessly debatable. Two major AI corporations, DeepMind and OpenAI, are explicitly aiming to create AGI. https://arxiv.org/abs/1705.08807 https://arxiv.org/abs/1705.08807, https://www.youtube.com/watch?v=HOJ1NVtlnyQ https://www.youtube.com/watch?v=HOJ1NVtlnyQ

astrange 6y ago
Although ML has definitely developed something, it has nothing to do with AGI, so they don't seem anywhere close. Instead it seems like we've gotten good at inventing visual cortexes. AGI would involve having some motivation to stay alive, affect the world and imagine things, instead of just using a fixed amount of CPU to point out whether something looks like a cat. From the article: > Unlike humans, CLIP can’t slow down to compensate for the harder task. Instead of taking a longer amount of time for the incongruent stimuli, it has a very high error rate.
- Veedrac 6y ago
  I like to keep personal opinions out of statistical data about opinions, since they are very different qualities of argument. Any individual take doesn't meaningfully affect the distribution of opinions. The key is that there is no argument from consensus to be made; smart people disagree, AGI within our lifetimes is not a niche expectation. Anyhow, I suspect you're not familiar with ML research, since eg. there are plenty of non-feedforward networks available. I also don't think much of this really matters; it is absurdly anthropocentric to expect ‘motivation to stay alive’ to be fundamental to intelligence.
  - nl 6y ago
    Name an actual active ML researcher who believes human level AGI is likely within our lifetime. Good researchers will never rule it out of course but I don't know any who think it is likely (I work in the field).
    - Veedrac 6y ago
      Demis Hassabis. https://www.youtube.com/watch?v=h0962biiZa4 https://www.youtube.com/watch?v=h0962biiZa4, older https://www.youtube.com/watch?v=Qgd3OK5DZWI https://www.youtube.com/watch?v=Qgd3OK5DZWI (NB ‘likely’ ≠ ‘>50%’, not sure what his percentage would be) I can give more but I did upstream link to a paper on this so it seems pretty pointless.
      - nl 6y ago
        Do you have a timestamp for a specific quote? I can point at "Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality" https://venturebeat.com/2018/12/17/geoffrey-hinton-and-demis-hassabis-agi-is-nowhere-close-to-being-a-reality/ https://venturebeat.com/2018/12/17/geoffrey-hinton-and-demis... and "Martin Ford interviewed 23 of the most prominent men and women who are working in AI today, including DeepMind CEO Demis Hassabis, Google AI Chief Jeff Dean, and Stanford AI director Fei-Fei Li. In an informal survey, Ford asked each of them to guess by which year there will be at least a 50 percent chance of AGI being built.... Ray Kurzweil, a futurist and director of engineering at Google, suggested that by 2029, there would be a 50 percent chance of AGI being built, and Rodney Brooks, roboticist and co-founder of iRobot, went for 2200. The rest of the guesses were scattered between these two extremes, with the average estimate being 2099 — 81 years from now." https://www.theverge.com/2018/11/27/18114362/ai-artificial-general-intelligence-when-achieved-martin-ford-book https://www.theverge.com/2018/11/27/18114362/ai-artificial-g...
        Veedrac 6y ago
        > Do you have a timestamp for a specific quote? I don't think there's a single timestamp for this since the exact question was never asked. But the gist is very explicit in his numerous talks (which note includes both the Future of Life Institute and Singularity Summit), and I think https://youtu.be/d-bvsJWmqlc?t=203 https://youtu.be/d-bvsJWmqlc?t=203 sums it up beyond reasonable doubt. > I can point at "Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality" This is a disaster of a headline, since while Hinton does, Hassabis never says that. > The rest of the guesses were scattered between these two extremes, with the average estimate being 2099 — 81 years from now. Which agrees with my point.
nl 6y ago
Name any researcher at either of those organisations who publically have stated human level AGI is possible within 50 years. It's an aspiration, but not something anyone believes on the horizon.
- nl 6y ago
  Notably the survey paper you linked is about outperformance in specific tasks. That's not AGI.
  - Veedrac 6y ago
    ‘Full Automation of Labor (All Human Jobs)’ pretty clearly encompases AGI.
    - nl 6y ago
      Not a single respondant thought "Full Automation of Labor" was possible in 50 years. The average prediction was nearly 125 years.
      - Veedrac 6y ago
        The bars are quartiles, so 25% of the probability mass is for ~53 years away or closer. Note that timelines for the HLMI variant of the question[1], which was asked to more people, had more like a 50% aggregate probability for 50 years away, per Figure 1. [1] “High-level machine intelligence” (HLMI) is achieved when unaided machines can accomplish every task better and more cheaply than human workers.

astrange 6y ago

Although ML has definitely developed something, it has nothing to do with AGI, so they don't seem anywhere close. Instead it seems like we've gotten good at inventing visual cortexes. AGI would involve having some motivation to stay alive, affect the world and imagine things, instead of just using a fixed amount of CPU to point out whether something looks like a cat. From the article: > Unlike humans, CLIP can’t slow down to compensate for the harder task. Instead of taking a longer amount of time for the incongruent stimuli, it has a very high error rate.

Veedrac 6y ago
I like to keep personal opinions out of statistical data about opinions, since they are very different qualities of argument. Any individual take doesn't meaningfully affect the distribution of opinions. The key is that there is no argument from consensus to be made; smart people disagree, AGI within our lifetimes is not a niche expectation. Anyhow, I suspect you're not familiar with ML research, since eg. there are plenty of non-feedforward networks available. I also don't think much of this really matters; it is absurdly anthropocentric to expect ‘motivation to stay alive’ to be fundamental to intelligence.
- nl 6y ago
  Name an actual active ML researcher who believes human level AGI is likely within our lifetime. Good researchers will never rule it out of course but I don't know any who think it is likely (I work in the field).
  - Veedrac 6y ago
    Demis Hassabis. https://www.youtube.com/watch?v=h0962biiZa4 https://www.youtube.com/watch?v=h0962biiZa4, older https://www.youtube.com/watch?v=Qgd3OK5DZWI https://www.youtube.com/watch?v=Qgd3OK5DZWI (NB ‘likely’ ≠ ‘>50%’, not sure what his percentage would be) I can give more but I did upstream link to a paper on this so it seems pretty pointless.
    - nl 6y ago
      Do you have a timestamp for a specific quote? I can point at "Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality" https://venturebeat.com/2018/12/17/geoffrey-hinton-and-demis-hassabis-agi-is-nowhere-close-to-being-a-reality/ https://venturebeat.com/2018/12/17/geoffrey-hinton-and-demis... and "Martin Ford interviewed 23 of the most prominent men and women who are working in AI today, including DeepMind CEO Demis Hassabis, Google AI Chief Jeff Dean, and Stanford AI director Fei-Fei Li. In an informal survey, Ford asked each of them to guess by which year there will be at least a 50 percent chance of AGI being built.... Ray Kurzweil, a futurist and director of engineering at Google, suggested that by 2029, there would be a 50 percent chance of AGI being built, and Rodney Brooks, roboticist and co-founder of iRobot, went for 2200. The rest of the guesses were scattered between these two extremes, with the average estimate being 2099 — 81 years from now." https://www.theverge.com/2018/11/27/18114362/ai-artificial-general-intelligence-when-achieved-martin-ford-book https://www.theverge.com/2018/11/27/18114362/ai-artificial-g...
      - Veedrac 6y ago
        > Do you have a timestamp for a specific quote? I don't think there's a single timestamp for this since the exact question was never asked. But the gist is very explicit in his numerous talks (which note includes both the Future of Life Institute and Singularity Summit), and I think https://youtu.be/d-bvsJWmqlc?t=203 https://youtu.be/d-bvsJWmqlc?t=203 sums it up beyond reasonable doubt. > I can point at "Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality" This is a disaster of a headline, since while Hinton does, Hassabis never says that. > The rest of the guesses were scattered between these two extremes, with the average estimate being 2099 — 81 years from now. Which agrees with my point.

nl 6y ago

Name any researcher at either of those organisations who publically have stated human level AGI is possible within 50 years. It's an aspiration, but not something anyone believes on the horizon.

nl 6y ago
Notably the survey paper you linked is about outperformance in specific tasks. That's not AGI.
- Veedrac 6y ago
  ‘Full Automation of Labor (All Human Jobs)’ pretty clearly encompases AGI.
  - nl 6y ago
    Not a single respondant thought "Full Automation of Labor" was possible in 50 years. The average prediction was nearly 125 years.
    - Veedrac 6y ago
      The bars are quartiles, so 25% of the probability mass is for ~53 years away or closer. Note that timelines for the HLMI variant of the question[1], which was asked to more people, had more like a 50% aggregate probability for 50 years away, per Figure 1. [1] “High-level machine intelligence” (HLMI) is achieved when unaided machines can accomplish every task better and more cheaply than human workers.