6 ms·

Climate change is always a risk but unlikely existential or at least to the extent it is we know how to handle it pretty well. So well that the number of climat

Climate change is always a risk but unlikely existential or at least to the extent it is we know how to handle it pretty well. So well that the number of climate related deaths have plummeted quite dramatically the last 100 years. We are constantly working on impacting nature in ways that make it safer for us whether we are talking building houses, using energy to keep warm or creating higher yield crop. AI isn't necessarily that different in that its probably not actually as dangerous in reality as in theory.

wizzwizz4 5y ago
AI is 98%-likelihood “benign” (i.e. just a tool, which is certainly not benign for those of us that it gets used against!) and 1%-likelihood world-ending catastrophe, like grey goo or nuclear armageddon. (Made-up percentages that don't even sum to 100%.) I don't think we have much to worry about for a while, because brains are pretty efficient, very few computers are as powerful as human brains, and there's a limit to the damage an evil human with internet access can do… isn't there? So I think AI risk is bounded by evil-human risk for the next ten years, at least. Climate change isn't a risk – there's still some uncertainty about it, but it's happening. Saying it's a risk is like saying riptides are a risk while you're in one. No! It's a danger! Get out of it by swimming perpendicularly to the shore!
- AbrahamParangi 5y ago
  The risk, presumably is the outcomes of climate change- which are uncertain and almost certainly not guaranteed death for everyone
  - wizzwizz4 5y ago
    “Guaranteed death for everyone” is the current state of affairs. You will die. I will die. This is a terrible bar for “should we care about this issue”? Autocratic dictator of the world says “deciduous trees and the colour yellow are now banned, on pain of death”? Well, it's not guaranteed death for everyone, so why should I care?
    - AbrahamParangi 5y ago
      I didn't say that you shouldn't care about it. I am saying that a 95th percentile negative outcome for climate change is still probably limited in its existential impact, not that you're gonna like it.
  - sammalloy 5y ago
    Wet bulb temperature, clathrate gun, food web collapse, albedo decrease, dormant virus release from permafrost, increase of tropical disease, water scarcity, crop and livestock failure—that /is/ almost certainly guaranteed extinction for humanity. How is this even debatable? We are already on this path.
    - wizzwizz4 5y ago
      But some groups of humans might survive, and might still have descendants tens of thousands of years later! It's not guaranteed that humanity goes extinct; we're pretty creative. (Though the same can't be said for a lot of other species.)
    - SpicyLemonZest 5y ago
      Well, expert organizations like the IPCC produce reports on the expected impact of climate change, which don't include human extinction as even a worst-case possibility. I'm the last person who'd claim the experts are always right, but it's gotta be at least debatable.
      - sammalloy 5y ago
        This problem was discussed by Pratt and Dunlop in their 2019 policy paper for the National Centre for Climate Restoration: > Climate scientists may err on the side of “least drama”, whose causes may include adherence to the scientific norms of restraint, objectivity and skepticism, and may underpredict or down-play future climate changes. In 2007, security analysts warned that, in the two previous decades, scientific predictions in the climate-change arena had consistently under-estimated the severity of what actually transpired. > This problem persists, notably in the work of the Intergovernmental Panel on Climate Change (IPCC), whose Assessment Reports exhibit a one-sided reliance on general climate models, which incorporate important climate processes, but do not include all of the processes that can contribute to system feedbacks, compound extreme events, and abrupt and/or irreversible changes. > Other forms of knowledge are downplayed, including paleoclimatology, expert advice, and semi-empirical models. IPCC reports present detailed, quantified, complex modelling results, but then briefly note more severe, non-linear, system-change possibilities in a descriptive, non-quantified form. Because policymakers and the media are often drawn to headline numbers, this approach results in less attention being given to the most devastating, difficult-to-quantify outcomes. > In one example, the IPCC’s Fifth Assessment Report in 2014 projected a sea-level rise of 0.55-0.82 metre by 2100, but said “levels above the likely range cannot be reliably evaluated”. By way of comparison, the higher of two US Department of Defence scenarios is a two-metre rise by 2100, and the “extreme” scenario developed by a number of US government agencies is 2.5 metres by 2100. > Another example is the recent IPCC 1.5°C report, which projected that warming would continue at the current rate of ~0.2°C per decade and reach the 1.5°C mark around 2040. However the 1.5°C boundary is likely to be passed in half that time, around 2030, and the 2°C boundary around 2045, due to accelerating anthropogenic emissions, decreased aerosol loading and changing ocean circulation conditions.

wizzwizz4 5y ago

AI is 98%-likelihood “benign” (i.e. just a tool, which is certainly not benign for those of us that it gets used against!) and 1%-likelihood world-ending catastrophe, like grey goo or nuclear armageddon. (Made-up percentages that don't even sum to 100%.) I don't think we have much to worry about for a while, because brains are pretty efficient, very few computers are as powerful as human brains, and there's a limit to the damage an evil human with internet access can do… isn't there? So I think AI risk is bounded by evil-human risk for the next ten years, at least. Climate change isn't a risk – there's still some uncertainty about it, but it's happening. Saying it's a risk is like saying riptides are a risk while you're in one. No! It's a danger! Get out of it by swimming perpendicularly to the shore!

AbrahamParangi 5y ago
The risk, presumably is the outcomes of climate change- which are uncertain and almost certainly not guaranteed death for everyone
- wizzwizz4 5y ago
  “Guaranteed death for everyone” is the current state of affairs. You will die. I will die. This is a terrible bar for “should we care about this issue”? Autocratic dictator of the world says “deciduous trees and the colour yellow are now banned, on pain of death”? Well, it's not guaranteed death for everyone, so why should I care?
  - AbrahamParangi 5y ago
    I didn't say that you shouldn't care about it. I am saying that a 95th percentile negative outcome for climate change is still probably limited in its existential impact, not that you're gonna like it.
- sammalloy 5y ago
  Wet bulb temperature, clathrate gun, food web collapse, albedo decrease, dormant virus release from permafrost, increase of tropical disease, water scarcity, crop and livestock failure—that /is/ almost certainly guaranteed extinction for humanity. How is this even debatable? We are already on this path.
  - wizzwizz4 5y ago
    But some groups of humans might survive, and might still have descendants tens of thousands of years later! It's not guaranteed that humanity goes extinct; we're pretty creative. (Though the same can't be said for a lot of other species.)
  - SpicyLemonZest 5y ago
    Well, expert organizations like the IPCC produce reports on the expected impact of climate change, which don't include human extinction as even a worst-case possibility. I'm the last person who'd claim the experts are always right, but it's gotta be at least debatable.
    - sammalloy 5y ago
      This problem was discussed by Pratt and Dunlop in their 2019 policy paper for the National Centre for Climate Restoration: > Climate scientists may err on the side of “least drama”, whose causes may include adherence to the scientific norms of restraint, objectivity and skepticism, and may underpredict or down-play future climate changes. In 2007, security analysts warned that, in the two previous decades, scientific predictions in the climate-change arena had consistently under-estimated the severity of what actually transpired. > This problem persists, notably in the work of the Intergovernmental Panel on Climate Change (IPCC), whose Assessment Reports exhibit a one-sided reliance on general climate models, which incorporate important climate processes, but do not include all of the processes that can contribute to system feedbacks, compound extreme events, and abrupt and/or irreversible changes. > Other forms of knowledge are downplayed, including paleoclimatology, expert advice, and semi-empirical models. IPCC reports present detailed, quantified, complex modelling results, but then briefly note more severe, non-linear, system-change possibilities in a descriptive, non-quantified form. Because policymakers and the media are often drawn to headline numbers, this approach results in less attention being given to the most devastating, difficult-to-quantify outcomes. > In one example, the IPCC’s Fifth Assessment Report in 2014 projected a sea-level rise of 0.55-0.82 metre by 2100, but said “levels above the likely range cannot be reliably evaluated”. By way of comparison, the higher of two US Department of Defence scenarios is a two-metre rise by 2100, and the “extreme” scenario developed by a number of US government agencies is 2.5 metres by 2100. > Another example is the recent IPCC 1.5°C report, which projected that warming would continue at the current rate of ~0.2°C per decade and reach the 1.5°C mark around 2040. However the 1.5°C boundary is likely to be passed in half that time, around 2030, and the 2°C boundary around 2045, due to accelerating anthropogenic emissions, decreased aerosol loading and changing ocean circulation conditions.