Now the demo is working and fast now.

I tried variations of the same code, changing the name of the `inner` function to `temp` and `banana`. In each of them I got a different explanation, for example:

> - The code is a function that takes in an input tensor and returns the squared difference between it and 0.5.

that I think it's more correct than "mean squared error". You should change that example.

I also got:

> - The code is called l2, which stands for "linear regression."

That is wrong.