The easy things were the hard ones

Ask someone to multiply 1,847 by 329. It feels difficult, but the steps can be written down.

Now ask them to recognize a friend from half a face, catch a falling cup, or cross a crowded room without colliding with anyone. These feel easy. Try writing down every rule involved.

Steven Pinker summarized a pattern AI researchers had discovered: "The hard problems are easy and the easy problems are hard." The observation is known as .

Psychologist Daniel Kahneman described a related contrast in human thought. He called fast, automatic thinking System 1 and slow, deliberate thinking System 2.

The match is not exact. Moravec's paradox is about what is difficult to engineer, not a claim that every automatic task defeats AI or every deliberate task is easy for it.

But the reversal matters. When people cannot state the hidden rules behind an ability, researchers need another way to build it. That leads to the next bet: learning from examples.

The easy things were the hard ones

Ask someone to multiply 1,847 by 329. It feels difficult, but the steps can be written down.

Now ask them to recognize a friend from half a face, catch a falling cup, or cross a crowded room without colliding with anyone. These feel easy. Try writing down every rule involved.

Steven Pinker summarized a pattern AI researchers had discovered: "The hard problems are easy and the easy problems are hard." The observation is known as .

Psychologist Daniel Kahneman described a related contrast in human thought. He called fast, automatic thinking System 1 and slow, deliberate thinking System 2.

The match is not exact. Moravec's paradox is about what is difficult to engineer, not a claim that every automatic task defeats AI or every deliberate task is easy for it.

But the reversal matters. When people cannot state the hidden rules behind an ability, researchers need another way to build it. That leads to the next bet: learning from examples.