Shape: The Hidden Geometry of Information, Biology, Strategy, Democracy, and EverythingElse

The paradox of education: what we most admire we put in a box and make dull.

The ultimate reason for teaching kids to write a proof is not that the world is full of proofs. It's that the world is full of non-proofs, and grown-ups need to know the difference. It's hard to settle for a non-proof once you've really familiarized yourself with the genuine article.

What I like about stochastic gradient descent is how nuts it sounds. Imagine, for instance, that the president of the United States made decisions without any kind of global strategy; rather, the nations chief executive is surrounded by a crowd of shouting subordinates, each hollering for policy to be tweaked in a way that suits their own particular interest. And the president, every day, chooses one of those people at random to listen to, and changes course accordingly. That would be a ridiculous way for a person to run major world government, but it works pretty well in machine learning!

Often people think of developments in computation as arising when we make our computers more blazingly fast, so they can compute more stuff, bigger data. It's actually just as important to prune away big parts of the data that aren't relevant to the problem at hand! The fastest computation is the one you don't do.

People often complain that no one likes facts and numbers and reason and science anymore, but as someone who talks about those things in public, I can tell you that's not true. People love numbers, and are impressed by them, sometimes more than they should be.

Geometry is the cilantro of math.

Between Jeans and Pearson comes a notice from one John Butler Burke, who believed he had observed spontaneous generation of microscopic life in a vat of beef bouillon by exposure to the recently discovered element radium.