@CassSunstein @KingsCollegeLon @PeterBoettke @timurkuran @duncanjwatts @SibOliv There is a deeper problem than lack of *data* or *information*, words which imply a universal & univocal domain of "the given". Domains which involving the theory-laden & skill-laden have have incommensurability components, eg Copernicus & Ptolemy share "data" but not categories.
@CassSunstein @KingsCollegeLon @PeterBoettke @timurkuran @duncanjwatts @SibOliv Hayek misleads us when he talks about the problem of knowledge coordination, plan agreement, and understanding compatibility as simply a lack of "data" or "information" problemβwe have also to coordinate our understanding and categories in a way captured by Wittgestein and Hume.
@CassSunstein @KingsCollegeLon @PeterBoettke @timurkuran @duncanjwatts @SibOliv When we talk about "data" and "information" we already assume an established social compatibility of concepts and ways of going on together set up in socially acquired shared practices and understandings as described by Wittgenstein. Algorithms depend on this shared background.
@CassSunstein @KingsCollegeLon @PeterBoettke @timurkuran @duncanjwatts @SibOliv Algorithms always ultimately interface with the background of shared/evolved human social practices arxiv.org
Loading suggestions...