> ML code can solve the problem without being programmed explicitly to do so
Umm, no. He implemented all of Fizz Buzz except the loop in his "fizz_buzz_encode" function. Totally explicit.
That’s the joke!
He implements it explicitly but only to generate training data.
That's just code to generate the examples, an example of synthetic data. He could just have written down the expected output instead of doing that.
> Most people understand Fizz Buzz, building up and solving the problem with a neural network shows that a bit of ML code can solve the problem without being programmed explicitly to do so, quite a good example.
If you train your model with completely different data than you use for inference, there is absolutely no guarantee that it will work as expected. In this case, they trained with numbers >= 1024, where numbers <= 100 were used for inference.
Most people understand Fizz Buzz, building up and solving the problem with a neural network shows that a bit of ML code can solve the problem without being programmed explicitly to do so, quite a good example.