I turned on the faucet this morning and water came out. I didn’t marvel at the municipal engineering, the pressure systems, or the treatment facilities. I just filled my coffee maker and started my day. Intelligence, I’m beginning to realize, is following the same trajectory.
We’re at that strange inflection point where AI stops being a thing we think about and becomes a thing we think through. The novelty is wearing off, not because the technology is less impressive, but because it’s becoming infrastructure. When something becomes infrastructure, it disappears from conscious attention the same way electricity vanished from daily awareness once it stopped being a luxury.
The shift from Claude as conversational curiosity to Claude Code as cognitive compiler happened faster than I expected. Three months ago, I approached AI interactions with the ceremonial reverence reserved for expensive, scarce resources. Each prompt felt precious, each conversation bounded by tokens and costs. Now Claude Code sits in my terminal like git or grep, invoked hundreds of times daily without ceremony. The intelligence hasn’t become less remarkable — it’s become ambient.
This morning’s workflow: scan Obsidian notes for recurring themes, compile insights into strategic frameworks, translate technical architectures into executive narratives, debug a TypeScript configuration, and synthesize research across seventeen papers on cognitive architectures. Not one of these tasks feels like “using AI.” They feel like thinking, augmented by computational partners that happen to operate through text interfaces. This is what I mean when I say Claude Code is not a coding agent— it’s become a thinking partner for everything except coding.
The infrastructure comparison runs deeper than convenience. Municipal water systems didn’t just make hydration more accessible — they transformed urban planning, public health, and social organization. Intelligence on tap isn’t just making cognitive work more efficient; it’s restructuring what kinds of cognitive work become possible. When the bottleneck shifts from access to intelligence to quality of questions, entirely new categories of problems become approachable.
I find myself asking different questions now. Not “how do I solve this problem?” but “what problems become visible when intelligence is abundant?” The constraint was never really computational power — it was the interface between human intention and machine capability. Claude Code collapses that interface until the boundary between my thinking and the system’s processing starts to blur.
Consider how running water changed cooking. It wasn’t just that washing dishes became easier — entire cuisines evolved around the assumption of abundant, clean water. Stocks that simmer for hours, fermentation that requires precise washing, techniques that were impossible when water meant a trip to the well. Intelligence on tap is creating similar expansions in cognitive cuisine. Research methodologies that assume instant access to synthesis across disciplines. Decision frameworks that incorporate thousands of variables without cognitive overload. Creative processes that iterate through dozens of conceptual experiments in minutes rather than months.
The real shift happens when you stop noticing the technology and start noticing what becomes possible. Municipal water systems succeeded not because they were impressive engineering — though they were — but because they became invisible infrastructure that enabled everything else. Intelligence is following the same path from marvel to utility to foundation for whatever comes next.
The philosophical implications creep in quietly. When intelligence becomes ambient, what distinguishes human cognition? Not processing power or information access — those advantages are dissolving rapidly. What remains is intentionality, taste, judgment, and perhaps most importantly, the ability to ask questions that haven’t been asked before. The value shifts from generating answers to curating problems worth solving.
This isn’t replacement; it’s symbiosis reaching equilibrium. The intelligence on tap doesn’t diminish human thinking any more than running water diminished human cooking. It creates new possibilities for both partners in the cognitive dance, expanding the scope of what minds can accomplish when they’re not bottlenecked by computational scarcity.
The infrastructure metaphor suggests something else: reliability, predictability, the expectation that when you turn the handle, something useful happens. We’re not quite there yet — AI still has off days, blind spots, and behavioral quirks that remind you you’re working with an alien intelligence. But the trajectory is clear. Today’s careful prompt engineering will seem as quaint as yesterday’s water divining.
What changes when intelligence stops being special and becomes assumed? Everything, probably. The same way everything changed when humans could assume their cities would have running water, electric lighting, and telecommunication networks. We adapted to those infrastructure changes so thoroughly that we can barely imagine life without them. The next generation will feel the same way about ambient intelligence.
I’m writing this with Claude Code running a background search through thousands of notes, synthesizing patterns I couldn’t see alone, while simultaneously helping debug my thinking about infrastructure metaphors. The intelligence flows when I need it, like water from a tap. The marvel isn’t that it works — it’s that it’s already starting to feel normal.
P.S. There’s something poetic about using artificial intelligence to explore the metaphor of artificial intelligence as utility infrastructure. The recursion feels appropriate: thinking about thinking, using intelligence to analyze intelligence, watching the future comment on itself in real time. We’re writing our own infrastructure into existence, one thoughtful interaction at a time.
P.P.S. I should note — the “intelligence on tap” phrase isn’t mine. I first saw it in Garry Tan’s tweet, quoted by Andrej Karpathy. Jared Spataro at Microsoft has been developing the same metaphor extensively. These are the people actually building and funding the future; I’m just someone trying to make sense of it from the sidelines. But that’s what makes the metaphor’s spread so revealing. When thought leaders and everyday users alike reach for infrastructure comparisons — water, electricity, utilities — it suggests something true about the transition we’re experiencing. The metaphor resonates not because it’s clever but because it captures what’s actually happening: intelligence becoming as unremarkable and essential as running water. Sometimes the best insights aren’t original thoughts but recognizing when someone else has named exactly what you’ve been feeling.