Moravec's Paradox is named after one of its coiners, Hans Moravec,

and it refers to the counterintuitive reality

that, up till this point at least, things that seem easy to humans can be incredibly

difficult and computationally expensive for artificial intelligences and robots, while many of the things

we find taxing are relatively simple for our synthetic creations.

So while AI software, especially in recent years, is just galloping across the expanse of human cognitive achievement

and bettering us across everything thinking related, ranging from memorization to math, to playing games, to negotiation. The same systems have trouble with casual language use,

figuring out what a conversation or joke is about,

and identifying stuff in photographs and videos,

things that in many cases are both intelligible and easy for even a human child.

Similarly, robots.

And the distinction here is that AI is typically software at this point,

while a robot is a physical thing that either we control or is controlled by some kind of software.

Robots can haul heavy loads and hit a target with a projectile every single time.

But walking, jumping, maintaining balance on precarious surfaces,

these are all things that humans are great at doing,

but which robots do not do terribly well.

And when they approximate decent performance,

it requires just a massive amount of work and processing on their part. One theory about why this is the

case is that natural selection, the force behind evolution, which basically says that those

entities that have survival-related advantages will tend to survive to procreate more, which

over the course of generations will help those that survive become even better at survival,

...