François CholletVerified account

Here's a simple example to capture the difference between a set of observations and an explanation. Consider Google search.

You can type up some keywords and write down what results you get. Then repeat this for thousands (or millions) of queries. At the end of the exercise, you'll have a description of a (tiny) slice of the behavior of the system.

But you will still have no idea that these results are produced via the PageRank algorithm. In fact, you won't have learned anything useful about the system at all. You won't be able to create your own search engine. You won't even know what makes a search engine good.

You will also have gained no context about what surrounds the system -- what search engines are for, or what kind of cultural, technological and economic forces shape the company that developed this one

You will have accumulated terabytes of data about the system, and yet you will still be infinitely far away from what you could learn from a 3-paragraph explanation written by someone who actually understands the system.

A 100% complete and accurate recording of brain activity during an entire lifetime -- or a 100% 1:1 brain simulation -- would be like recording millions of search results for different queries. A description, not an explanation.

And not conducive to building a new system that isn't just a faded Xerox of the original. (some people did try to build new search engines that way -- didn't really work out)