You make an important point: Data as observations, as the proceeds of observing (those “infinite options” and “subtle choices” you speak of). Data is a process, not an object. The implication is that observations are necessarily “theory laden.”

The idea that “you can make data anytime, anywhere” is not widely appreciated. (Data is more frequently viewed as a passive resource to be mined, a latent source of knowledge.) The insight that explanations make data is harder still.

And so the disciplines of “making data useful” are deeply explanatory in their means. I think it’s valid to praise the tool makers, whether telescopes or statistics or ML. But I wonder whether enough attention is given to those who decide where to point the tools and why. Galileo didn’t invent the telescope, but he showed how to use it. More importantly, he showed why to use it.

I like the connection here to the “memory revolution.” But it’s data in the context of a scientific revolution, in a process of criticism, that points to its true worth.