Language models are trained by a simple trick.

We give them a line of text and ask them, “What comes next?”

We do this over and over again, and eventually, they learn to complete the line.

By this trick language models have learned to write poetry, prose, and quantum physics. It works because in text, as in life, what’s past is prologue.

Perhaps we shouldn’t be surprised that asking a machine to answer such a simple question produces such powerful results. Because “What comes next?” is the question that has been driving us forward since the beginning of history.

It is the fundamental question of science: It’s what caused the ancients to invent astronomy and Newton to invent calculus. It is the question that every religion has tried to answer, in every age, through myth and ritual.

It is what animates a parent watching a child struggling to take their first step or a sports fan watching the fourth quarter. It’s why you binge-watch every episode of your favorite TV show, and why you can’t put down a new book.

“What comes next?” is the question that we try to answer at Every, every day. Our answers to this question become the essays that we write, the podcasts we record, the courses we teach, and the software products that we build.

We ask “What comes next?” not because we intend to find definitive answers. Our work is never done because our answers are never final. Instead, our answers create new questions; the past informs the future.

We write for you, because you care about what comes next, too. But we don’t presume to give you the answers. Instead, we hope our answers inspire you to ask and answer this question yourself—so that you can think new thoughts and dream new dreams that become new technologies and new businesses.

The truth is that by answering this question, over and over, we make what comes next, together.