Noah Smith challenges a dominant assumption in tech circles: that artificial intelligence will inevitably fracture corporations into armies of independent contractors. Instead, he posits a startling counter-intuition—that the very technology promising to make everyone their own boss might actually drive us back toward massive, centralized firms. This is not a story about job loss alone; it is a structural argument about how trust, verification costs, and the nature of digital fraud could reshape the entire architecture of the American economy.
The Solopreneur Surge
Smith begins by grounding his analysis in recent data that contradicts pre-pandemic fears of stagnation. He notes that business creation didn't just recover after 2020; it accelerated, with a significant rise in "solopreneurs"—individuals running businesses without employees. Citing Stripe Economics, he observes that this trend was already visible before AI took center stage. "The age of the solopreneur... has been increasing since 2008," Smith writes, noting how tax codes and digital platforms like Substack have lowered barriers to entry.
He argues that AI acts as a force multiplier for these individuals, filling the skill gaps that previously necessitated hiring teams. "Claude is far more versatile than any human being, so in the age of AI agents, specialization will probably be less important in many cases," he asserts. This framing aligns with recent findings that "AI-native" companies are already operating with roughly 25% fewer employees than their peers. The logic seems sound: if an algorithm can handle marketing, coding, and legal compliance, why hire a department?
The availability of this breadth of on-tap assistance allows anyone with sufficient motivation to go it alone.
However, Smith quickly pivots from this optimistic vision to the economic theory that underpins corporate structure. He reminds us that companies exist not just to produce goods, but to minimize "transaction costs"—the friction and expense of dealing with outsiders. Historically, as he notes with a nod to Ronald Coase's 1937 work, it was cheaper to keep payroll in-house than to hunt for contractors via phone books. The internet flipped this script by making verification easy, fueling the outsourcing boom that defined the early 2000s.
The Trust Deficit
The core of Smith's argument lies in his prediction that AI will reverse the outsourcing trend by inflating transaction costs again. While the internet made it cheap to verify a human contractor's reputation, AI introduces a new variable: synthetic fraud. He points out that "an AI can create a new company, give it a website, and invent a bunch of other companies and reviewers to make it look legit." This isn't theoretical; he cites reports of unscrupulous firms using AI-generated images to pose as family-run businesses.
If every potential partner in the global marketplace is potentially an AI hallucination or a fraudster, the cost of verification skyrockets. Smith draws a sharp parallel to cryptocurrency: "In order to do away with the need for intermediaries like banks to verify transactions, Bitcoin has to reestablish trust between counterparties every time a transaction is performed." He argues that this constant re-verification is so expensive it doomed Bitcoin as a medium of exchange, and the same logic could apply to AI-driven business.
Critics might argue that verification tools will evolve alongside fraud, creating an arms race where detection becomes cheap again. But Smith counters that even without malice, AI agents lack the consistency of human character. "Human beings are relatively consistent over time... This is not generally true with AI," he writes. If a contractor's goals and capabilities shift unpredictably from week to week, long-term arm's-length contracts become too risky.
If only big companies can establish internal trust cheaply, then many humans — whatever kind of work they're still doing 10 or 30 years from now — will be working for big companies.
This leads to a fascinating historical echo. Smith suggests we might see a return to "Japan-style 'salarymen' occupations," where workers perform shifting bundles of tasks within a single massive entity. This mirrors the dynamics seen in places like Castro Valley, California, where local economic shifts have often been driven by large institutional anchors rather than agile startups. The implication is that the future of work may not be a gig economy, but a "salaryman" economy on steroids.
A Bifurcated Future
Smith does not dismiss the solopreneur trend entirely. He acknowledges that in sectors where trust is less critical or fraud is easily detected, independent workers will thrive. But he predicts a "bifurcated distribution of company sizes and job types." We could end up with a vast horde of lonely solopreneurs competing on price, alongside a few "monster companies" employing huge numbers of wage earners to manage the complex, high-trust operations that AI cannot safely outsource.
This reframing moves the conversation away from simple automation anxiety. It suggests that the structural incentives of the market might actually favor consolidation in an age of intelligent agents. As Smith puts it, "Internal procedures for verifying trustworthiness — the AI equivalent of walking down the hall and knocking on the door of your own internal corporate division — might be the new equivalent of the informal long-term relationships that lower transaction costs within a typical human-staffed company."
Bottom Line
Noah Smith's most compelling insight is that AI's ability to generate content and impersonate entities could paradoxically make large, hierarchical organizations more efficient than decentralized networks. The strongest part of his argument is the application of transaction cost theory to synthetic media; it provides a rigorous economic reason why the "gig economy" might stall. His biggest vulnerability lies in assuming that verification costs will remain prohibitively high indefinitely, as detection technology may eventually outpace generation capabilities. Readers should watch for early signals of whether AI-native startups are actually shrinking or if they are merely masking traditional corporate structures behind new tech stacks.