The Verification Crisis: Why Trust Is the New Bottleneck in AI
Noah Smith
Hugo Bowne-Anderson
Duncan Gilchrist
Noah Smith, economist and author of Noahpinion, joins High Signal to look at what AI is already changing… and what it isn't. The conversation moves beyond the usual productivity hype to ask harder questions: Is agentic coding actually increasing revenue per hour worked? Will software remain a high-margin business if AI makes it easy to clone? And what happens when generating content, code, vendors, applications, and companies becomes much cheaper than verifying them? Noah argues that one of the biggest near-term shifts is not simply automation, but trust. AI is beginning to replace parts of the internet's knowledge infrastructure — search, Stack Overflow, Reddit, and how-to content — while also flooding markets with new forms of slop. For AI builders and leaders, the central challenge may become less about producing more and more about knowing what is real, valuable, and worth trusting…. In a word, verification.
Guest
Noah Smith
Economist and Author at Noahpinion
Key Takeaways
Generation is scaling faster than verification.
AI makes it cheap to produce code, content, applications, resumes, vendors, and companies. But checking quality, trustworthiness, and intent remains expensive. For leaders, the scarce capability has become judgment and verification rather than production.
The slopocalypse creates new attack surfaces.
Fake AI-generated companies, vendors, resumes, and outsourcing firms could poison markets. Noah's phrase "outsource bombing" captures the risk that malicious actors create fake service providers to win competitors' business and make them fail.
Developer speed is not the same as productivity growth.
Agentic coding may make builders feel much faster, and may reduce hiring needs, but the economic question is whether it leads to shipped products, revenue, and durable value.
Software may become abundant and low-margin.
Low switching costs, open-source competition, model distillation, and AI-enabled cloning could make software look less like a winner-take-all monopoly market and more like farming, steel, solar, or airlines: huge consumer value, intense competition, and compressed profits.
Producer identity may give way to consumer identity.
Noah argues that work often reflects market constraint more than personal expression: jobs are what the market pays you to do. Consumption, leisure, creative choices, and even accepting lower pay for meaningful work may become more central to identity as AI automates more production.
The scariest AI risk is malicious humans with super-tools.
Noah is less worried about a self-improving machine god than about a "vibe-coded super virus": a jailbroken agent helping someone design dangerous pathogens and order them through poorly governed wet labs before society has adequate biosecurity.
AI may satisfy material wants while intensifying status games.
If humans are near saturation on many non-rival goods, future demand may flow into relative status: prestige, attention, aesthetics, location, and social competition. AI can amplify both sides of those tournaments without necessarily making people happier.
Higher individual productivity does not guarantee smaller firms.
Manufacturing history suggests the opposite can happen: when individual workers became more productive, firms often got larger because coordination and transaction costs mattered more. AI could similarly produce both giant trust-based firms and one-person businesses.
AI is eating the internet's knowledge layer.
Chatbots are not just new productivity tools; they are replacing search, YouTube how-tos, Reddit, Stack Overflow, Math Overflow, and other systems people used to navigate knowledge online. This may reshape online communities before it shows up in GDP.
The best career advice is tool fluency, not prediction.
Noah refuses confident forecasts about which careers will win. His practical advice is to use AI tools aggressively, learn their strengths and failure modes, and become familiar enough to recognize opportunities when they appear.