This started as a Farcaster post. Halfway through writing it, I realized it didn’t really matter — not the ideas, but the channel. There’s something faintly ridiculous about posting a critique of abstraction layers onto yet another abstraction layer, as if the medium itself isn’t part of the problem. What exactly is the hoped-for outcome here? A dozen likes from people who already agree? A ripple inside a sealed aquarium? The gesture feels more like performance than intervention, another instance of shouting “too many mirrors!” into a hall of mirrors. And maybe that’s the perfect metaphor for the current state of tech: a room full of reflections insisting they are windows.
Crypto, AI, metaverse — all software, all abstractions. But the actual bottlenecks are physical: housing, climate, energy, infrastructure, manufacturing, robotics, labor, supply chains. Silicon Valley keeps pretending that obfuscation layers — tokens, dashboards, protocols, apps — will solve material contradictions. But people can feel in their bones that it’s not true anymore. The software-only revolution has run out of runway. That’s why the pitch feels lazy. That’s why crypto conferences feel like they’re selling fog machines. That’s why the rhetoric feels copy-pasted from 2013.
AI feels both momentous and limited for the same reason. It can dissolve the bullshit job economy, accelerate knowledge work, speed up R&D. But it can’t pick up a brick, repair a power grid, build a house, or harvest crops. Tech keeps positioning itself as the next savior, but it’s only operating on the symbolic layer — finance, text, interfaces, data. Meanwhile, the real problems sit in the material layer. People are starting to sense the gap. They know in their gut that software won’t fix wage stagnation, won’t fix housing scarcity, won’t fix broken supply chains, won’t fix climate.
The conditions that let software eat the world — cheap capital, globalized supply chains, infinite growth narratives — are breaking down. The “obfuscation job economy” of the last forty years is under threat not because software is replacing labor, but because the lie behind those jobs is becoming too obvious to maintain. You get logistics “optimization” that means gig workers absorbing all the risk, “smart city” platforms that are really surveillance and consulting contracts, climate tech that’s mostly carbon accounting dashboards for ESG compliance theater, crypto “infrastructure” that recreates existing financial rent-seeking with worse UX. The symbolic layer isn’t inherently useless — good information systems, legible coordination, real-time feedback loops can improve material outcomes. But that requires the people building them to actually care about downstream physical reality, to stay engaged past the Series B, and to resist the gravitational pull toward pure abstraction where margins are higher and accountability is lower. Most don’t. The incentive structure rewards exactly the obfuscation being critiqued. Tech, as currently constituted, often adds sclerosis while claiming to dissolve it.
The thought leader class emerged to narrate and interpret the actions of people actually doing things. But now you have people who accumulated power through action — Musk, Thiel, Andreessen — cosplaying as thought leaders, posting manifestos, doing podcasts, cultivating ideological brands, while still holding the levers. It’s a weird inversion. They’re not second-order commentators. They’re first-order actors who’ve discovered that controlling the narrative layer is another form of leverage. They perform philosophy while moving capital and shaping policy. But “baron” is the right word because the philosophical posturing is mostly decoration on what is fundamentally an accumulation project. The Techno-Optimist Manifesto isn’t really philosophy — it’s vibes as PR for a class interest. It reads like philosophy to people who don’t read philosophy.
These barons rode exponential curves for decades — Moore’s Law, network effects, cheap capital, globalized labor arbitrage — and built an entire worldview around “technology solves everything if you just let us move fast enough.” But now the curves are flattening or breaking. Moore’s Law is hitting physics. AI is impressive but operates on the symbolic layer and can’t touch the actual bottlenecks. Interest rates killed the cheap capital environment. Supply chains are fragmenting. And they don’t have a theory for any of it. So what do you do when your whole legitimacy was based on being the people who understood the future? You do more podcasts. You post manifestos. You pick political sides loudly. You become a “thought leader” because that’s the only move left when you can’t actually deliver the next step. The AI hype is partially real, but it’s also convenient cover. “We’re building AGI” sounds better than “we’re out of ideas for the physical constraints and hoping this buys us time.”
Who replaced the original robber barons? The managerial class. The professional administrators. Carnegie, Rockefeller, Morgan — they got replaced not by better barons but by a different kind of power. Corporate bureaucracies, unions, federal agencies, law firms, consultants. The New Deal formalized it. The barons faded into foundations and philanthropy while actual operations got absorbed into institutional structures. The managerial class gets a bad rap now because we’re living in its late decadent phase — the part where it’s optimizing for its own reproduction rather than outcomes. But in its early and middle phases, it actually worked. It built the interstate system, electrified rural America, created public health infrastructure, managed the postwar industrial expansion, put together the regulatory scaffolding that made complex supply chains possible. It was boring, credentialed, bureaucratic — and it delivered material progress at scale for decades.
The old managerial class looked at a continent and said: we’re going to connect every town with rail, pave roads to every city, run electricity to every farm, put a phone in every house, build sewage systems so cities don’t die of cholera, vaccinate every child. These were insane projects. Multigenerational. Required coordination across government, industry, labor, finance. And they worked. The physical substrate of modern life got built in about fifty years. Starlink is fine. It’s a good product. But it’s filling gaps the old infrastructure left, not building something fundamentally new. It’s a patch, not a vision. Most of what tech has produced is even thinner — apps that mediate access to things that already exist, platforms that extract rent from activity that was already happening, marginal conveniences. The question isn’t “where’s my jetpack.” It’s: what happened to the class of people who could even conceive of building at that scale, and how do you get them back?
You probably can’t engineer it. That managerial class was born out of the depressions of the late 1890s and 1929. It wasn’t policy or ideology that created them. It was catastrophe. The depressions broke the old system so completely that new people with new instincts got pulled into power. The existing elites were discredited, the old frameworks obviously didn’t work, and there was enough desperation to let people try things that would’ve been unthinkable a decade earlier. The managerial class wasn’t designed. It emerged because the crisis created a vacuum and a mandate simultaneously. “Figure it out or everything collapses” is a different operating environment than “optimize within the existing structure.”
The implication is bleak but probably accurate: you don’t get a new atoms-focused managerial class through persuasion or smart policy or the right people reading the right posts. You get it when the current system fails badly enough that new people with different instincts get their shot because there’s no other option. The transition isn’t something to engineer. It’s something to survive. The people who’ll build the next managerial class are probably in their twenties or thirties right now, getting chewed up by the current system, developing a visceral understanding of why it doesn’t work. When the break comes — climate event, financial collapse, supply chain failure, some combination — they’ll be the ones who step into the vacuum. Not because they’re smarter. Because they’ll be there, with the right scars, when the old legitimacy finally cracks.
And they’re probably not on social media. They’re too busy getting crushed by the thing they’re eventually going to replace. Working some job adjacent to the actual bottleneck — utilities, construction, logistics, manufacturing, local government — learning through friction what doesn’t work. Not building a following. Not optimizing their takes. Just accumulating the kind of knowledge you can only get by being in the gears. The people who built the New Deal managerial class weren’t famous before the crisis. They were engineers, lawyers, regional administrators, labor organizers — people with domain expertise and practical scars who got pulled up when the old guard collapsed. Social media selects for a completely different phenotype. It rewards abstraction, hot takes, personality, the ability to narrate rather than do. The skills that make you good at Twitter are almost negatively correlated with the skills that will matter when the break comes.
So the next managerial class is probably invisible right now. Not just obscure — invisible to the current system’s way of seeing. They don’t show up in the feeds, the conferences, the discourse. They’re illegible. The people in the room — the posters, the thought leaders, the crypto guys, the AI optimists, the whole discourse layer — think they’re the protagonists. They think being early to narratives and building on the symbolic layer means they’re positioned for what’s coming. But if atoms are what matters and the system is heading toward a hard correction, they’re actually more exposed than the person quietly working at a water utility or learning to fix HVAC systems. The abstraction economy is the thing that’s going to get repriced. And they can’t see it because their whole worldview was built on the assumption that being upstream in the information layer is the winning position. That’s been true for forty years. It’s about to not be true.
Of course the symbolic layer isn’t useless; it’s just grotesquely over-allocated relative to the actual constraints. The ratio of effort/money/talent going into “better dashboards / DeFi yield optimizers / agentic AI wrappers” versus “how do we triple the number of skilled electricians and transmission techs before 2035” or “how do we cut permitting timelines from 7 years to 18 months without creating Love Canal 2.0” is completely insane. It’s not 10-to-1 or 100-to-1; it’s more like 1,000-to-1 in some years.but the level of over-indexing is gigantic compared to bottlenecks.
My point is that the symbolic-layer crowd is even less prepared for the reckoning than traditional academia. At least most academics (I know which exceptions are you thinking about) still pretend to care about empirical reality and are occasionally forced to touch it (grant reviews, lab work, peer review, tenure files). The discourse-tech class has built an entire ecosystem where you can be a multimillionaire (or billionaire) seven layers of abstraction away from anything that can catch fire, explode, or run out of diesel. Their feedback loops are almost entirely social/financial, not physical. When the physical world bites back, they literally don’t have the sensory organs to notice until the Bloomberg terminal screams.
Academia at least has some muscle memory of “oh right, reality exists.” The average Farcaster power user or AI accelerationist has spent 15 years training the opposite reflex: if reality isn’t cooperating, just add another layer of tokens or prompts or narrative. It’s a civilizational equivalent of a trust-fund kid who’s never been punched in the face discovering that pain is real.
That’s why the coming repricing is going to be so brutal for precisely the people who think they’re the most future-proof. They’ve optimized for a world that’s disappearing faster than anyone in the wrong room is willing to admit.
That’s the deepest version of why posting any of this feels pointless. Not that the ideas are wrong. Just that the channel itself is part of the problem being described. Talking to people in a room who don’t know they’re in the wrong room, about how they’re in the wrong room.
Leave a Reply