If intelligence becomes abundant, inexpensive, and increasingly non-human, which theories of civilization still hold โ and which must be rewritten?
Rather than studying AI purely as a technical discipline, the goal is to understand it through multiple lenses simultaneously. The premise is that AI is likely to become a civilizational technology โ not merely another software platform. Understanding its long-term implications therefore requires synthesis rather than specialization.
1.Build a coherent mental model for how AI changes civilization.
2.Understand how previous technological revolutions transformed societies and institutions.
3.Develop intuition for second- and third-order effects rather than focusing only on near-term disruption.
4.Better understand the interaction between incentives, institutions, technology, culture, and governance.
5.Identify which historical theories remain predictive in an age of abundant machine intelligence.
6.Recognize where existing academic frameworks appear insufficient.
7.Form my own multidisciplinary perspective rather than adopting a single ideological or disciplinary lens.
Central Questions
>What happens when intelligence becomes infrastructure?
>Which institutions become more valuable, and which become obsolete?
>Does AI strengthen states or networks?
>Does it concentrate power or decentralize it?
>What becomes scarce once intelligence is abundant?
>How should ethics evolve when intelligence is no longer uniquely human?
>Which historical analogies are genuinely useful?
>Which assumptions about economics, education, labor, governance, and culture need to be revisited?
An interactive map of the syllabus. The guiding question is at the centre, the seven parts ring it, and each reading connects to its part. Move through the parts and readings with the Tab key and open one with Enter. With a reading in focus, the left and right arrow keys move between the readings of that part and the up arrow moves to the part itself; with a part in focus, left and right move around the ring of parts and the down arrow moves into that part's first reading. Drag to pan, and pinch or hold control while scrolling to zoom; full screen, a two-finger scroll pans as well. Use the full screen button to open the map at the size of the window, which is easier to read on a phone. Escape clears the current selection, then leaves full screen. The same information is available as a list โ use the List view button above the map โ and as Markdown at /syllabus.md.
Select a part or a reading on the map to see its details. The centre node is the question the whole syllabus is built around: What happens when intelligence becomes infrastructure?
Part I โ Civilization & History(3 readings)
The big structural theories of where civilizations are headed โ and whether any of them survive contact with abundant machine intelligence.
Can civilizational instability be modeled structurally?
Part II โ Media & Cognition(3 readings)
How the medium reshapes the mind that uses it. Read with 'television' and 'the internet' mentally swapped for LLMs, synthetic media, and algorithmic feeds.
The post-scarcity argument in its strongest form, and the only reading here that treats abundant machine intelligence as a reason to replace the market rather than to model it. Read against "What becomes scarce once intelligence is abundant?"
Part IV โ Networks, States & Power(4 readings)
Does AI strengthen states or networks? Does it concentrate power or distribute it? The historical record on hierarchies, networks, and institutions.
The pessimistic case for what AI does to states: public discourse and government function increasingly run by machinery and privately owned corporations, not by consent.
Part V โ AI Futures(5 readings)
The forecasters and the frontier. Mostly published on the open web, and mostly revised faster than books can be printed.
The optimistic case, from the frontier lab's own vantage point: what a decade of powerful AI could do for biology, economics, and governance if the risks are managed well.
The companion risk essay: bioterrorism, autonomous weapons, mass labor displacement, and AI-enabled authoritarianism โ and what it would take to defend against them.
Part VI โ Ethics, Society & Governance(5 readings)
How ethics and governance evolve when intelligence is no longer uniquely human โ and what that does to the moral psychology underneath.
What becomes scarce once intelligence is abundant? Bostrom's answer is purpose โ the philosophical counterpart to Bastani's economic case, and roughly where Machines of Loving Grace leaves off.
Part VII โ Information & Civilization(1 reading)
Information networks as the substrate of civilization โ worth revisiting periodically as AI develops.
The book I suspect has not yet been written sits somewhere between Neil Postman, Marshall McLuhan, Francis Fukuyama, Tyler Cowen, Yuval Noah Harari, and frontier AI researchers. Its central question would be:
What happens when intelligence becomes infrastructure?
Electricity democratized energy.
The printing press democratized knowledge.
The internet democratized information.
Large language models may democratize reasoning itself.
That shift has implications for economics, politics, science, education, warfare, religion, identity, creativity, governance, and culture. Understanding that transition โ not merely the technology itself โ is the overarching objective of this syllabus.