Pedro Domingos

I don't know about you, but I infinitely prefer working with Gaussians than say very large binomial distributions, even on computers. Same goes for difference equations vs differential equations. Or Laplace transforms vs z-transforms.

I broadly agree that there might be an overemphasis on continuous mathematics vs discrete, but I disagree that it is not useful in the right circumstances.

Yes, nothing is continuous. But it can be a good approx., and this can give us insight. An example I always use is the M/G/1 queue, where one finds the mean wait depends not only on mean job time but also the variance of job time. Aimless discrete wandering wouldn't tell us that.

Tbh I'm not sure I'd be confident saying the world *is* discrete or continuous. And even then, we don't model fluid dynamics by telling the computer where the atoms are -- modelling the discrete directly. We make a discrete approximation to a continuous model...

The idea that continuous math is a hindrance is absurd. Even if one assumes limitless computing power, we still need to understand phenomena in way to feed it to computers.