Anatomy of an AI System (2018)

Reading this I constantly find myself asking, so what? It feels like each little vignette ends before managing to substantiate its point, consider the concluding sentence in the early section on lithium mining:<p>&gt; Your smart-phone runs on the tears and breast milk of a volcano. This landscape is connected to everywhere on the planet via the phones in our pockets; linked to each of us by invisible threads of commerce, science, politics and power.<p>This purple language is evocative of, something? Like ya, we live in a globalized economy, things are connected, what of it? Is this supposed to be a brilliant insight or just a poetic restating of the obvious?

I can&#x27;t speak on the technical side of things, but that graphic is beautiful. Might even frame and hang it up.

I only got partway through this. I want to say to the author:<p>Oh my god just stop taking yourself so seriously. Really. You don&#x27;t need to do this. You&#x27;re trying waaaay too hard.<p>It&#x27;s opaque and mysterious and lovecraftian <i>to you</i> because you clearly spend all of your time making cool graphics and drawing tenuous metaphors between 16th century monks, pharaohs, and jeff bezos. I assure you, for everyone who works in a lithium mine or amazon datacenter doing the actual thing, it&#x27;s actually pretty obvious. you just, you don&#x27;t need to write this maam.<p>And also google &quot;commodity fetishism&quot; sometime.<p>It does look cool though. Wonderful typography and illustrations. I have a nitpick which is that it&#x27;s a little bit monochrome, a pop of color (maybe a gold? or like neon red) would really take it to the next level.

The author wrote that book, which I found quite interesting: <a href="https:&#x2F;&#x2F;yalebooks.yale.edu&#x2F;book&#x2F;9780300209570&#x2F;atlas-ai" rel="nofollow">https:&#x2F;&#x2F;yalebooks.yale.edu&#x2F;book&#x2F;9780300209570&#x2F;atlas-ai</a><p>In particular, in this book there is a discussion about the dangers and epistemic distortions induced by the &quot;supervised learning&quot; paradigm in machine learning, which imposes rigid systems of classification on reality, which was a novel and refreshing angle for me.

The graphic is wonderful -- wide ranging and mind expanding. Think of it as a jumping off point in connecting sociotechnical systems around &quot;AI&quot;.<p>Also, for context around the project as a whole, you need to read the accompanying narrative (scroll down).

Really enjoyed poking around the graphic and diving into all the interlocking systems. Really didn&#x27;t enjoy the prose below it as much. Rather than adding to it, it seemed to drown out other ways of consuming the information.<p>I think I learned a lesson about how a heavy-handed POV can push the audience right past the ideal zone of understanding. Or maybe I didn&#x27;t learn that lesson, and I will continue to write primarily to show off how insightful my insights are.

From the text:<p>&gt; Our exploded view diagram combines and visualizes three central, extractive processes that are required to run a large-scale artificial intelligence system: material resources, human labor, and data.<p>Also:<p>&gt; Sandro Mezzadra and Brett Nielson use the term ‘extractivism’ to name the relationship between different forms of extractive operations in contemporary capitalism, which we see repeated in the context of the AI industry. 10 There are deep interconnections between the literal hollowing out of the materials of the earth and biosphere, and the data capture and monetization of human practices of communication and sociality in AI.<p>What are the arguments both -for- and -against- a claim that &quot;data&quot; is an extractive process as used here?