Robin Hanson and Brink Lindsey talk fertility on 22Jan2023

แชร์
ฝัง
  • เผยแพร่เมื่อ 14 ม.ค. 2025

ความคิดเห็น • 11

  • @newtype0083
    @newtype0083 11 หลายเดือนก่อน +3

    There's no way you'll get the same innovation from a whole different population. People and cultures are different and not interchangeable.

  • @grahamlawlor8361
    @grahamlawlor8361 11 หลายเดือนก่อน +2

    What about a govt policy that incentivized these traditional religious communities (Amish, Mormon, Orthodox Jewish, etc) to move immigrate? I suppose it's been happening in the US since colonial times, inadvertently. I'm sure there must be similar communities in China, Russia, etc who are being opressed and would love the opportunity

  • @kreek22
    @kreek22 11 หลายเดือนก่อน

    I don't find the claim that innovation "grinds to a halt" convincing or accurate phrasing. This is not how one accurately summarizes the notion that the rate of innovation declines in proportion to the decline in population size. An economy at half the size innovates, other things being equal, half as much. Nor does it make much sense to claim that low fertility will eventually extinguish the species, which is the limit case in which innovation truly does halt. Now, one could claim very reasonably that innovation rates decline somewhat faster, due to loss of scale, than the rate at which the population size declines. But, even a far smaller economy could plausibly continue some types of innovation, even as it lost some tech knowledge.
    My main objection, though, to this entire line of thinking, this worry about population decline, is that AI is likely to change the economic and social world more drastically in the next 30 years than the change seen in any other 30 year period in world history. This makes the medium term future radically unpredictable. I rate it as very unlikely that in 2050 people with Hanson's general perspective still consider fertility to be a first order problem. I suspect it's more likely that no one will be around to worry about anything. But, the most likely play in which we still exist is a world heavily infused with AI labor and AI assistants, along with many AI-trained robots. This is a world of more leisure time, at least for those who are not career or status obsessed. Having children might come to seem like a fun or interesting project for such people.

    • @filmmakerdanielclements
      @filmmakerdanielclements 11 หลายเดือนก่อน +3

      If you think the medium term future is radically unpredictable, why suspect anything is likely or unlikely? Doesn't radically unpredictable mean that, either way, you can't fathom what is likely?

    • @kreek22
      @kreek22 11 หลายเดือนก่อน

      @@filmmakerdanielclements Yes, contingent upon AI becoming a major factor in human affairs, which is a position contrary to Hanson's. He says that the many factors driving fertility down are likely to continue until the low fertility cultures extinguish themselves and high fertility cultures replace them. I think all of those factors are highly dependent on gradual cultural evolution, which AI seems likely to do chaos upon.

    • @derek4412
      @derek4412 2 หลายเดือนก่อน

      Have you seen the cost of artificial intelligence? The energy usage alone is likely to extinguish it. The promise of an AI-produced future might just be the dying breath of the religion of Tech Optimism.

    • @kreek22
      @kreek22 2 หลายเดือนก่อน

      @@derek4412 Yes, I have. Here are some numbers:
      Electricity represents ~22% of global energy usage. In 2023, 1.2% of global electricity production went to data centers (350 TWh out of 28,000 TWh). Most of the data center usage is *not* for AI. One study showed that generative AI will account for ~110 TWh of usage in 2027, which is 0.4% of global electricity usage and 0.1% of total global energy usage. That same study estimates 2023 usage at ~6 TWh. Another study estimated that data centers' total usage in 2026, including AI, will be ~840 TWh.
      The AI boom only continues if if pays for itself. If it fails to add value to the economy, the vast investments currently planned will be curtailed. Considering that 1 TWh of electricity in the US costs about $120 million, they'll need at least $13 billion/year in income by 2027 just to pay for the energy costs. The costs of building their specialized data centers are much, much higher than that (around 10x higher).

  • @Wardoon
    @Wardoon 11 หลายเดือนก่อน

    I am 10 minutes into the video but still I can't say if it is a debate, a discussion, an interview because Brink keeps interrupting Robin and won't let him complete a thought. I don't know Brink but he is the worst interviewer, debate partner and BAD LISTENER. Sorry to say that.

    • @Wardoon
      @Wardoon 11 หลายเดือนก่อน

      I am hoping it gets better as I listen on

    • @elektrochava
      @elektrochava 11 หลายเดือนก่อน +2

      thought it was pretty good@@Wardoon