The Stack On Software and Sovereignty · 第 185 页

By contrast, looking at Frank Gehry ’ s early proposals for a new Facebook headquarters in Menlo Park (nicknamed “ Zee Town ” after company founder, Mark Zuckerberg) we see a plan for a more traditional corporate campus, designed, it appears, to ensure the managed serendipitous contact between employees in motion. In this encapsulated “ company town ” winding pathways and strategic lines of sight connecting interior and exterior views are embedded in a multilevel landscape where sub- and superterranean greenery twists and turns onto and under the collection of buildings. ⁵⁹ At their desks, the aggregate social graph of the on-site employee/resident population is framed and displayed to itself as it moves and involves itself within itself in airplane hangar – scale open-plan work space. It ’ s an obvious but valid observation to note that the collected body of Facebook employees is here mediated by Gehry ’ s proposal as a closed-loop network of people, one that will ostensibly perform at a higher level should its undulating nodes and edges be given a single supportive program. Unlike the old Googleplex, defined by robust services in a generic, no-frills context (like Google perhaps), Facebook ’ s seems more tuned to the advantages of the active performance and the structured monetization of captured human networks (like Facebook itself). For its part, Amazon recently took over a huge high-rise campus in the South Lake Union neighborhood of Seattle, and there it will help to consolidate its scattered executives into one big storefront. ⁶⁰ Plans for another nearby location also include a “ biosphere ” building featuring three huge glass domes in which full size trees and other flora will grow in simulated ecologies. ⁶¹ For Amazon to encapsulate ecosystems of nonhumans is perhaps the point. Together, the urban headquarters will integrate the company into the fabric of a “ real city ” in ways that the suburban Silicon Valley campuses will not. Amazon ’ s society will have comparatively blurry boundaries between itself and the rest of the world, similar perhaps to how the mega-retail platform is a more agnostic medium of supply, demand, and algorithmic recommendation, defined by behind-the-scenes engagement with its outside suppliers and Users.

This is fine and well, but the far more important architectural-urban footprint of the Amazon Cloud platform is not in Seattle but distributed among the company ’ s many fulfillment centers, especially in and around the logistics plantations near the airports of Louisville, Kentucky (UPS ’ s hub) and Memphis, Tennessee (FedEx ’ s hub). ⁶² Amazon ’ s platform logic is based on the massive coordination of the pricing, retail display, storage, and delivery of its flat commerce ontology of objects. If Google ’ s mission is to organize the world ’ s information, Amazon ’ s may be to organize the world ’ s tangible commodities. This places it on a direct collision course with Walmart, but that company has been slow to move into e-commerce and still largely uses its supply chain omniscience to guarantee itself economies of scale with a limited range of downmarket goods — a very short-tail strategy — sold through its network of grim retail centers to people who may not have so many other options. ⁶³ Amazon uses the physical supply chain itself (especially Cloud infrastructure, airports, warehouses, and third-party delivery services) in lieu of any big box retail holdings. These networks, taken together as a composite Amazonian territory, are the platform ’ s megastructural play at the City layer, all but invisible to its Users save for the vast Amazon.com website. One can go inside those object-network machines and find unsentimental beauty in their airportadjacent logistics so precise and responsive that they could be described as a form of artificial intelligence. That description will become more true as Amazon ’ s warehouses, sorting, and distribution facilities become further populated by robotic systems and “ workers. ” ⁶⁴ We might anticipate that in a few years, an Amazon platform User could choose an item online and have it retrieved (even manufactured), mailed, and delivered with no human touch, at least until the last postal mile. This may be welcome or terrifying news to the precarious population of current Amazon warehouse workers. As discussed in the Cloud chapter, many are nomadic part-timers coming and going with the ebbs and flow of retail demand. Those with relatively permanent positions, called Amazon fulfillment associates, are divided among those on the receive lines, the pack lines, stockers, and pickers who are directed by handheld devices to find your order wherever it may be among the immediate stocks of children ’ s bicycles, shoe repair kits, and physics textbooks. By all accounts, Amazon space is already built on the nimble precision of a logistical engineering of human workers ’ movements with a repetitious efficiency probably better suited to robots. In describing the stress and precariousness of work in Amazon fulfillment centers, GigaOM, a Bay Area technology blog, went so far as to characterize employment at Amazon as a “ dystopian model of neofeudalism. ” ⁶⁵ As Amazon absorbs, centralizes, and consolidates production labor into tighter strata of proprietary commerce-logistics algorithms, the future of work is made that much more uncertain, and along with it the buying power of the workers who would also be their customer- Users. ⁶⁶

City Layer

185