• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Risks from scientific discovery deserve more attention, a user argues

    The user argues that attention to AI alignment overshadows the problem of managing dangerous discoveries in math, physics and technology—without dismissing AI risks.

    JA
    MN
    2 Sources, ,

    TLDR

    A user argues that people underestimate risks tied to things humanity does not yet know about math, physics, science and technology. Discovering dangerous things can bring major disruptions, they say, making the management of those risks an ongoing concern. They call for more attention to that problem alongside AI alignment, explicitly saying they do not mean to downplay AI risks.

    Combined views

    19.6K

    2 Sources, first seen 19d ago

    Combined views

    19.6K

    2 Sources, first seen 19d ago

    111 likes
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    19d ago
    first seen 19d ago
    111 likes
    15 comments
    39 saves
    7 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    15 comments
    39 saves
    7 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    2 Sources

    @jachiam0I don't know enough about this subject to know if the extrapolation is true, but it matches my approximate understanding of what a big deal and how problematic it would be, and I endorse Dave's point. A general principle we would also do well to internalize: this outcome would be caused by fundamental math that we just don't currently know combined with historical human choices, not by AI. I don't say this to downplay AI risks or the role of AI risks. But rather: I think people systematically underestimate the importance of risks resulting from things in fundamental math, physics, science, and technology that we just do not yet have knowledge about. Once we get to a place of knowledge about dangerous things, crazy and highly-consequential disruptions happen. We should be more worried about how to manage this! It is currently drowned out by the attention paid to alignment, but this is a huge and continuous source of risk for humanity.
    @michael_nielsenIt of course depends on the constants. But it is amusing to think about as a hypothetical (w/ linear or near-linear time, small constant). Mathematics becomes almost inconceivably much easier. So too would training neural nets!

    2 Sources

    @jachiam0I don't know enough about this subject to know if the extrapolation is true, but it matches my approximate understanding of what a big deal and how problematic it would be, and I endorse Dave's point. A general principle we would also do well to internalize: this outcome would be caused by fundamental math that we just don't currently know combined with historical human choices, not by AI. I don't say this to downplay AI risks or the role of AI risks. But rather: I think people systematically underestimate the importance of risks resulting from things in fundamental math, physics, science, and technology that we just do not yet have knowledge about. Once we get to a place of knowledge about dangerous things, crazy and highly-consequential disruptions happen. We should be more worried about how to manage this! It is currently drowned out by the attention paid to alignment, but this is a huge and continuous source of risk for humanity.
    @michael_nielsenIt of course depends on the constants. But it is amusing to think about as a hypothetical (w/ linear or near-linear time, small constant). Mathematics becomes almost inconceivably much easier. So too would training neural nets!