Don't worry about the name

The name "ReLU" is a mouthful. Forget it.

Here's the one thing worth carrying forward: deep networks had a problem where the training signal kept fading before it reached the early layers. A simple change to how individual neurons worked — letting the signal pass through without shrinking it — fixed that.

A decades-long obstacle. A very small solution.

That's the pattern you'll keep seeing: big problems in AI often have simple-sounding fixes, once someone finds the right angle.