10 October 2017

Backchannel: “The Myth of a Superhuman AI”

There is no doubt that a super AI can accelerate the process of science. We can make computer simulations of atoms or cells and we can keep speeding them up by many factors, but two issues limit the usefulness of simulations in obtaining instant progress. First, simulations and models can only be faster than their subjects because they leave something out. That is the nature of a model or simulation. Also worth noting: The testing, vetting and proving of those models also has to take place in calendar time to match the rate of their subjects. The testing of ground truth can’t be sped up.


To be useful, artificial intelligences have to be embodied in the world, and that world will often set their pace of innovations. Without conducting experiments, building prototypes, having failures, and engaging in reality, an intelligence can have thoughts but not results. There won’t be instant discoveries the minute, hour, day or year a so-called “smarter-than-human” AI appears. Certainly the rate of discovery will be significantly accelerated by AI advances, in part because alien-ish AI will ask questions no human would ask, but even a vastly powerful (compared to us) intelligence doesn’t mean instant progress. Problems need far more than just intelligence to be solved.

Kevin Kelly

I’ve written about this in the past, but I think it’s worth repeating: AI doesn’t exist somewhere outside the physical restrictions of the real world, nor will it magically solve all human problems. And this article does a wonderful job of exposing and taking apart five assumptions most people make when discussing AI – well worth reading!

Radial tree of life
Every one of these species has undergone an unbroken chain of three billion years of successful reproduction, which means that bacteria and cockroaches today are as highly evolved as humans. There is no ladder. Likewise, there is no ladder of intelligence.

Post a Comment