Oct 24, 2023·edited Oct 24, 2023Liked by Ivan Vendrov
This starts with a great point, but I think you're accidentally straw-manning a bit. While it's true that there are countless ways that using only data as your epistemology can go awry (goodhart's law, mcnamara fallacy, etc.), so too are there countless ways that using only qualitative expertise can go awry.
Reject both extremes! You can't make great decisions without data. You can't make great decisions without qualitative expertise.
In short: "data driven" shouldn't even be an ideology. It's a *method*, when used well, in pursuit of a set of values (e.g. an ideology, for example, though candidly I think anything that's an ideology is a bug-ridden, half-implemented value system)
This starts with a great point, but I think you're accidentally straw-manning a bit. While it's true that there are countless ways that using only data as your epistemology can go awry (goodhart's law, mcnamara fallacy, etc.), so too are there countless ways that using only qualitative expertise can go awry.
Reject both extremes! You can't make great decisions without data. You can't make great decisions without qualitative expertise.
In short: "data driven" shouldn't even be an ideology. It's a *method*, when used well, in pursuit of a set of values (e.g. an ideology, for example, though candidly I think anything that's an ideology is a bug-ridden, half-implemented value system)
Dan Davies' notion of Accountability Sinks seems useful for this discussion