Hacker News Reader

Moravec's paradox - stuff we thought is hard (chess, derivatives) is easy, while stuff we think is easy (face recognition, getting out of a car, opening a door) is really hard.

https://en.wikipedia.org/wiki/Moravec%27s_paradox

One of my favourite videos is this compilation of robots falling (failing) during the simplest tasks at the 2015 DARPA challenge:

https://www.youtube.com/watch?v=g0TaYhjpOfo

View on HN

To be fair, it takes us humans a solid couple of years of continuous training to make sense of our sensory inputs and get accustomed to our limbs enough to begin doing anything significant with them, so I am regularly flabbergasted at the expectations we place in our lowly contraptions to outdo us by two to three orders of magnitude.

And yet humans need even more years to correctly play chess.

Humans have had millions of years of system level optimisation (evolution) which then only needs a couple of additional years to train the wetware to operate the system for daily activity.

Chess has only been around a mere 1,500 years and has had an evolutionary impact on only a very minute portion of the population. Hence it takes a few additional years to specialise a human for optimal chess performance.

Optimal chess performance is not remotely achievable by humans.

Reply on news.ycombinator.com