4Joined Jan 2022


Some thoughts on Toby Ord’s existential risk estimates

I'm afraid Toby underestimates the risks from climate change in the following ways:

  1. He mentions in passing but does not really integrate the implications of the fact that a couple of feedback loops are around methane (CH4) rather than carbon dioxide (CO2). 1,500 gigatons of methane clathrates for example would be not just "more" than all anthropogenic emissions to date. It would be anything from 50 times to 150 times more (depending which IPCC GWP factor you use). The same is true for Siberian permafrost - whatever percentage you put as CO2 release, any significant faction of this released as CH4 will quickly hit a multiple of total anthropogenic. Such proportions render his side remark about methane potency being dependent on timeline moot.
  2. He does not factor in knock-on feedback loop scenarios. Specialists believe some feedback loops may already have kicked in - notably Arctic sea ice and the Greenland glaciers. Toby is right that these are lesser order loops in and of themselves. But they make the other methane-based feedback loops much more likely, especially given the now uncontested much faster rate of warming in the Arctic than the planetary average. So there is more chance of reaching a civilisation ending feedback loop indirectly, through a lesser loop, than directly through anthropogenic emissions.
  3. Existential risk is defined too fungibly. He seems at times to use existential risk and unrecoverable collapse of civilisation interchangeably. But there are very different probabilities in terms of climate change effecting one or the other. The end of the human race remains a distant outlier, as he concludes. But the end of industrial-based civilisation is between one or two orders of magnitude higher. He states, for instance, that even at average temperature rises of over 13 degrees, or up to 20 degrees, "there would remain large areas in which humanity and civilisation could continue". That also of course goes (implicitly) with the impact of climate change as being capable of being summarised by the label on the tin (average temperature rise) without considering second order effects both in local ecosystems and anthropogenically.
  4. These sub-effects are mentioned, but no attempt made to assess the probability and impact of cumulative effect of more than one of them... how does civilisation fare with the collapse of the Gulf Stream AND the loss of the albedo AND massive desertification around the Equator AND and so on and so on. The process seems to be to judge the potential impact of each in isolation, decide they fall beneath a threshold, and therefore discount them. There is no network effect.
  5. There seems to a rational actor fallacy at work. I say this because Toby says we are concerned merely with direct existential risk in climate change because if there are no direct risks, or they are trivial, they have no impact on other risks. But that is doubly incorrect. First, the "rational" impact of climate change on a central belt of the earth, including several nuclear powers, could easily increase risk of existential conflict significantly. Second, we would have to assume that all actors in and around anthropogenic risks were capable of accurately assessing those risks, as we, the readers of the book do, and acting on them. That is not realistic.

All the Best


Johnny West