Interesting article: Datacenters as core infrastructure of urban AI
He basically argues that the so-called 'cloud' comes with serious environmental impacts and externalities. Sure, we all want smart cities (aka data-driven urbanism) managed by Urban AIand love to talk about the cool edge-computing devices, but there is a hidden infrastructure (often in the suburbs & beyond) that make all this possible and that raises serious questions related to ownership. governance, and operation.
Here, physical distance or proximity often becomes irrelevant and with that many of our existing regulatory system fail.
Cloud providers want a) cheap electricity, b) low taxes, c) lax regulations, and d) proximity to high-speed connecticity. Thus, clusters develop along metropolitan corridors and once again regions farther away are left behind. Furthermore, AI providers obviously prefer to serve profitable sectors of the economy as opposed to common good public functions.
For cities this also means an increasing dependence on a small number of vendors who control the infrastructure & access to it. We are 'locked-into' whatever we choose.
In summary: Urban AI or any AI for that matter means infrastructure & calling it 'the cloud' hides the fact that this infrastructure and access to it creates power & privilege.
