One huge difference between the forest example and the society example is that we know forest ecosystems evolved over millions of years to be at least somewhat robust to perturbations. Likewise for individual plants. If you can arrange for their external conditions to at least roughly approximate the conditions they evolved into, they’ll probably be fine.
We have no such guarantee for anything whatsoever in human society. We have no precedents that can tell us whether “the network” is even capable of correcting any problems to maintain any decent quality of life without intensive management. History gives some mild hope for optimism, but also a great number of examples of catastrophic failure. Similar concerns apply to AI neural networks, and essentially everything else that humanity has been involved with in the past tens of thousands of years.
One huge difference between the forest example and the society example is that we know forest ecosystems evolved over millions of years to be at least somewhat robust to perturbations. Likewise for individual plants. If you can arrange for their external conditions to at least roughly approximate the conditions they evolved into, they’ll probably be fine.
We have no such guarantee for anything whatsoever in human society. We have no precedents that can tell us whether “the network” is even capable of correcting any problems to maintain any decent quality of life without intensive management. History gives some mild hope for optimism, but also a great number of examples of catastrophic failure. Similar concerns apply to AI neural networks, and essentially everything else that humanity has been involved with in the past tens of thousands of years.
This article didn’t provide evidence that it is possible, it just defined the intended approach.
I agree that the analogy with the plants or the forests is not trivial.
However, even though civilization is short lived, human communities did evolve for much longer.
Also, large scale systems like the market seem very robust (it survived wars, as well as environmental and technological perturbations).
So the idea here is conditional:
If we did find a root cause for moloch dynamics and catastrophic failures, then adressing it MIGHT allow society to solve problems by itself.
Think of it like a wild bet we should at least put some chips in. I’ll write more on this soon.