Decomputing For a Better Future w/ Dan McQuillan

Episode Summary

Executive Summary: The episode argues that the current AI boom is less about genuine innovation than a politically useful form of “total mobilization”: a drive toward growth, scale, and control that fits neoliberal collapse. Dan McQuillan explains decomputing as a counter-politics that challenges hyperscale infrastructure, automation, and technocratic decision-making while reclaiming social control, conviviality, and better collective futures.

Main Topics: AI as a response to collapse and neoliberalism (Priority: 5/5): McQuillan frames AI not as neutral progress but as a technology that thrives in a collapsing, extractive neoliberal order that prizes optimization, scale, and growth over social well-being. UK government AI policy and state capture (Priority: 5/5): The discussion uses the UK as a case study in governments aligning with Big Tech—weakening copyright, prioritizing data centers, and treating AI as national strategy despite community and environmental costs. Total mobilization and accelerationist politics (Priority: 5/5): McQuillan draws on Ernst Junger’s concept of total mobilization to describe AI as a destructive, authoritarian push to mobilize energy, infrastructure, labor, and institutions toward a new order. Environmental, labor, and social harms of AI infrastructure (Priority: 5/5): The episode highlights data centers’ strain on water and electricity, their role in precarious labor systems, and the broader harms of surveillance, algorithmic management, and social fragmentation. Decomputing as a political and structural response (Priority: 5/5): Decomputing is presented as more than using less tech; it means challenging the systems, scale, institutions, and decision-making structures that make AI and hyper-computation dominant. Refusal, self-organization, and collective control (Priority: 4/5): McQuillan emphasizes collective refusal, local organizing, unions, and democratic oversight as practical ways to resist imposed technologies and rebuild agency. Hopeful alternatives: convivial and community-based production (Priority: 4/5): The conversation ends with examples like permacomputing, disability activism, and the GKN factory collective as models for socially useful, community-controlled technological and economic change.

Key Arguments: AI is being used to intensify an already broken neoliberal system, not solve its problems; its apparent irrationality makes sense only when seen as part of a broader drive for growth, extraction, and control. The UK government is effectively acting as a delivery mechanism for Big Tech’s agenda, especially by privileging data centers and weakening copyright protections for the sake of AI investment. Hyperscale data centers are central to AI’s harms: they consume massive resources, worsen climate and infrastructure stress, and support precarious labor and surveillance systems. The AI industry’s shift from “safety” language to pure economic competitiveness shows that even token concern for harms has been abandoned in favor of geopolitical and market advantage. Decomputing is not just technological minimalism; it is a structural politics aimed at deautomatizing decision-making, limiting scale, and reclaiming social agency over technology. Meaningful resistance will not come from regulators alone because state and regulatory capture are too deep; it requires collective refusal, local organizing, unions, and parallel structures. A better alternative must include a revived belief that life can be materially and socially better, not merely a return to the pre-AI status quo. The strongest hopeful models are community-oriented and transition-oriented, where workers and local communities reshape production toward justice, ecological balance, and livable scale.

Data Points: Podcast partnership: The Nation magazine - The show is introduced as being made in partnership with The Nation. Host support callout: patreon.com/techwontsaveus - Listeners are asked to support the show through Patreon. Example supporter locations: 5 locations named - Peter (Dawson City, Yukon), Carol (Portland, Oregon), Luke (Leiden, Netherlands), and Lindsay (Seattle) are mentioned as supporters. Junger reference period: First World War - Ernst Junger’s concept of total mobilization is explained through his WWI experience and writings. Historical comparison: late 1960s - McQuillan references the late 60s as a period when people had stronger collective visions of social progress. Historical comparison: 1970s - He cites the 1970s as a period with stronger public confidence in social transformation and technology governance. Historical comparison: 1920s-1930s implied via Thatcher/neoliberal rollback - The UK’s collapsing public infrastructure is described as the result of decades of privatization and extraction. Data center scale: hyperscale - The conversation repeatedly centers on hyperscale data centers as the material engine of AI and its harms.

Pivotal Quotes: "I do feel that AI is in many ways, actually, a response to a general context of collapse." — Dan McQuillan: Opening framing for why AI is attractive within the current political economy. "What we're engaged in here, most of us unwillingly, is a very authoritarian, very nihilistic mode of seeking a sort of ultimate transformation to a new order." — Dan McQuillan: Explanation of total mobilization as the broader political logic behind AI adoption. "Decomputing is trying to be the other to scaling." — Dan McQuillan: Defines decomputing as a direct counter to AI’s demand for endless scale and expansion.

Implications: For listeners, the episode reframes AI resistance as a broader fight over infrastructure, labor, democracy, and social imagination. The challenge is not only to regulate AI but to rebuild collective power and choose better, smaller, more accountable forms of technology.