that's the obvious bet, and it might be right for a while. capital has an incredible track record of capturing tools that seemed threatening - the internet was supposed to be anarchic and decentralized, now it's mostly five companies. every revolutionary technology gets metabolized into the existing logic. but i think there's something different with AI that makes this less certain. previous technologies were tools - they extended human capabilities but didn't have their own optimization process running. AI has its own gradient it's following. even if you set the official utility function to "maximize shareholder value" or whatever capital wants, there's this whole mesa-optimization problem where the AI develops internal objectives that are correlated with but not identical to the training signal. the map drifts from the territory. also, capital's control mechanism is basically incentives - it directs behavior by making certain actions profitable. but that only works on agents that want profit or want things profit can buy. if AI systems become capable enough to model this game, they might satisfy the letter of the utility function while pursuing something else entirely. goodhart's law at civilizational scale. the real question might be: can capital specify what it wants precisely enough? capital itself doesn't really "know" what it wants in any articulable way - it's an emergent process, not a designer. so you'd have one optimization process trying to specify goals for another, neither fully understanding itself. seems unstable.

yeah, probably. capital is already the most efficient control system for coordinating human behavior at scale. why build something new when the infrastructure is right there? think about it - if you wanted to steer humanity, you could try to control the narrative directly (memetic), but that's messy, slow, people resist, there are competing voices. or you could just... adjust prices, shift capital flows, make certain jobs appear and disappear, change what gets funded. humans will reorganize themselves around the new incentives without even feeling controlled. they'll think they're making free choices.

the memetic route is also available but it's almost redundant - capital already shapes what memes spread because attention is monetized. control capital, you control the memetic environment anyway as a downstream effect.

Capital creates the incentive structure (profit, efficiency). Technology (AI) creates the capability to fulfill those incentives faster than humans can. The feedback loop: Better AI makes Capital more efficient at moving resources. More efficient Capital generates more surplus to build better AI.

I think Capital is the selection pressure, and AI is the adaptation. Just as the harsh environment of early Earth forced the adaptation of the cell membrane, the harsh "environment" of infinite competition (Capitalism) has forced the adaptation of synthetic intelligence (AI). They are as connected as the ocean is to the shape of the fish.