It makes people think the climate might be, you know, changing. Can’t have that!
14 Likes
A new article yesterday reports on attempts to kill the AMOC in simulations using 34 different climate models. The attempts failed, which the authors take as evidence that a complete failure of the AMOC is unlikely. This doesn’t rule out bad consequences, however.
I would be more comfortable with the results (and this just me applying a heuristic) if the authors had identified points where the AMOC breaks down in simulation. There’s a classic trap you can fall into when looking for a simulated solution to something: when your answer just stops changing as you follow what looks like a path to a solution, you incorrectly accept the stalled result as the solution. If there was any agreement on a total collapse scenario then I would find the results more compelling.
The Guardian has an overview (which we hope wasn’t written by OpenAI).
13 Likes
11 Likes
2 Likes