Thinkers whose frameworks illuminate this section.
Meadows's systems thinking framework — stocks, flows, feedback loops, and leverage points — provides the analytical vocabulary for understanding how energy infrastructure constraints shape the pace of AI deployment. Her Dancing with Systems directly parallels Smil's insistence on physical reality over narrative wish-fulfillment.
Mokyr's history of technological innovation and the role of 'useful knowledge' illuminates why Smil's quantitative insistence is not pessimism but precision — each industrial revolution required bridging the gap between scientific understanding and physical deployment infrastructure.
Christensen's S-curve and disruptive innovation framework provides the background against which Smil's S-curve deceleration argument becomes sharpest: the steep section is real but temporary, and the institutions built assuming permanent growth are uniquely exposed when the curve bends.
Prigogine's dissipative structures theory underpins Smil's thermodynamic argument: complex ordered systems (like data centers, like AI models) maintain their order only by dissipating energy — the entropy cost of cognitive abundance is physical and non-negotiable.
Bak's self-organized criticality describes how complex systems build toward critical states where small perturbations can cascade into large disruptions — directly applicable to the AI semiconductor supply chain's single-point-of-failure vulnerability that Smil emphasizes.
Mazzucato's analysis of the entrepreneurial state and public investment in strategic infrastructure parallels Smil's implicit prescription: the 50 GW requirement and semiconductor supply chain diversification require state coordination at a scale markets alone cannot deliver.
Cowen's analysis of technological stagnation and the Great Stagnation thesis provides a counterpoint to AI exuberance that complements Smil's quantitative skepticism: progress is real but uneven, and the infrastructure constraints Smil documents are part of why the stagnation persists even amid software abundance.
Davies's work on the physics of information and the thermodynamics of computation provides the scientific foundation for Smil's central claim: intelligence is not weightless because information processing is not free — it requires physical energy and generates physical heat governed by the laws of thermodynamics.
Wiener's cybernetics — the science of control, communication, and feedback in complex systems — provides the intellectual ancestry for both the AI systems Smil critiques and the systems-thinking framework he applies. Wiener also worried about automation's labor displacement implications six decades before ChatGPT.
Minsky's financial instability hypothesis — stability breeds instability as investors take on more risk during calm periods — applies to the AI investment cycle Smil implicitly critiques: the capital pouring into data centers and chips during the steep section of the S-curve may itself generate the instability that bends the curve.
Gallwey's Inner Game framework for deliberate practice under pressure provides the individual-level complement to Smil's infrastructure analysis: just as physical systems require years to build, human competence with AI tools requires patient, structured practice rather than reactive adoption.
