Cloud computing platforms rely on what are simply called ‘containers’. Computing platforms are increasingly organized around software ‘containers’ such as Docker (Docker 2022). Would it be possible to better ground platforms in their ensemble-natures by paying attention to containers and their construction? Numbers offer one way of following the container construction.

There are two problems in attending to platform numbers. The first, which I term the numericity problem, concerns the different ways in which platforms gather things through many kinds of calculation, algorithms, addressing techniques and statistical routines such as machine learning and the like. By speaking of ‘numbering’, I hope to resist treating numbers as if I know what they are. If there is any starting point, it is more like differences and their differencing. Numberings differ. They are multiple and they multiply. Numberings twist multiples together in different ways that can be traced through numeric, calculative and computational practices. Numberings occasion a wide variety of relational problems, including who gets to decide what numbering is done, and how that numbering makes worlds. Platforms overflow with numbers organized in lists, arrays, vectors, matrices, tables and parameters.

The second problem-group concerns a few large numbers and their affects on platform plurality. I call this few-many asymmetry the platform-number problem. There are in non-trivial respects few platforms (for example, Meta Platforms Inc, Alphabet, Apple, Amazon, TenCent, and so on). These few have ‘big numbers’.