> ## Content Index > Fetch the complete content index at: https://wire.fourthweb.ai/llms.txt > Use this file to discover other available public pages before exploring further. # Anthropic's Chief Economist Admits AI Can't Replace Workers While Cuban Says It Never Will - URL: https://wire.fourthweb.ai/anthropics-chief-economist-admits-ai-cant-replace-workers-while-cuban-says-it-never-will/ - Published: 2026-07-23T12:26:42.000Z - Updated: 2026-07-24T04:30:50.000Z - Description: The economist building AI admits it can't replace you yet, while the billionaire betting on it says it never will. Anthropic's head of economics says AI hasn't moved unemployment rates despite automation fears, with current models working best as augmentation tools for skilled users - Author: Travis Wright - Tags: AI Agent Economy, Agentic Workflows, AI Infrastructure, Anthropic, Microsoft **The economist building AI admits it can't replace you yet, while the billionaire betting on it says it never will.** ### The Summary - [Anthropic's head of economics says AI hasn't moved unemployment rates despite automation fears](https://www.businessinsider.com/anthropic-economist-claude-mark-cuban-ai-unemployment-jobs-skills-work-2026-7?ref=wire.fourthweb.ai), with current models working best as augmentation tools for skilled users - Mark Cuban argues AI lacks two critical human capabilities: understanding consequences ("what gets you fired") and real-time situational awareness ("reading the room") - AI remains skill-biased, meaning expert humans extract more value from it than novices — the opposite of replacement dynamics ### The Signal Peter McCrory's essay lands differently when you remember [he works for Anthropic](https://www.businessinsider.com/anthropic-economist-claude-mark-cuban-ai-unemployment-jobs-skills-work-2026-7?ref=wire.fourthweb.ai), the company behind Claude. This isn't doomposting from an outside economist. This is the person charged with understanding AI's economic impact, from inside one of the three companies building frontier models, saying: not yet. His framing matters. "Skill-biased and labor augmenting" means two things. First, better operators get better results. A junior analyst and a senior analyst using Claude get wildly different outputs. Second, it's making people more productive, not making them obsolete. That's the Web2 playbook. That's Excel and Google Docs, not the robot apocalypse. > "The tasks that Claude can't handle may depend on interpersonal coordination, in-person interactions, or engagement with the physical world that, so far, only humans can do." But here's the tension: McCrory hedges. He says this holds "for at least the next year." That's a tight window for confident predictions from someone with access to [Anthropic](https://wire.fourthweb.ai/tag/anthropic/)'s roadmap. Either he knows something about Claude 6's capabilities, or he's being appropriately humble about emergence. Neither interpretation is comforting. Cuban's response cuts deeper. He names two specific human advantages: consequences and presence. "People know what will get them fired" isn't about intelligence. It's about skin in the game. An AI can hallucinate a strategy that torches a client relationship. It doesn't feel the fallout. A human account manager thinks three moves ahead because her mortgage depends on it. "Information latency" is Cuban's term for situational awareness. You walk into a conference room and instantly read tension, exhaustion, enthusiasm. You adjust your pitch mid-sentence based on a board member's face. [Cuban says AI won't have this "for a long, long, long time. If ever."](https://www.businessinsider.com/anthropic-economist-claude-mark-cuban-ai-unemployment-jobs-skills-work-2026-7?ref=wire.fourthweb.ai) Key differences between AI and human judgment: - AI optimizes for the prompt. Humans optimize for survival. - AI processes what you tell it. Humans process what you didn't say. - AI learns from training data. Humans learn from getting burned. What's striking is that both McCrory and Cuban are AI bulls. Neither is a Luddite. McCrory builds economic models for an AI company. Cuban invests in AI startups. And both are saying: the replacement thesis is overblown. This maps to what we're seeing in Web4\. The most valuable applications aren't [autonomous agents](https://wire.fourthweb.ai/tag/ai-agents/) replacing workers. They're copilots, assistants, and tools that make skilled humans more leveraged. GitHub [Copilot](https://wire.fourthweb.ai/tag/microsoft/) doesn't replace developers. It makes senior developers write code faster and helps junior developers learn patterns. ### The Implication If the economist at Anthropic and a billionaire tech investor both say AI augments rather than replaces, the smart money is on building for augmentation. That means designing agents that assume human judgment in the loop, not trying to remove it. For workers, this is quietly optimistic. The premium shifts to skills AI can't replicate: consequence modeling, real-time pattern recognition, interpersonal calibration. Those aren't mystical human traits. They're learnable. But they're also the opposite of the rote, repetitive tasks we assumed were safe. Watch who's hiring in 2026\. Companies building AI-native workflows aren't slashing headcount. They're looking for operators who can wield these tools at scale. The question isn't whether AI takes your job. It's whether you're the person who knows how to make AI do what it can't yet do alone. ### Sources [Business Insider Tech](https://www.businessinsider.com/anthropic-economist-claude-mark-cuban-ai-unemployment-jobs-skills-work-2026-7?ref=wire.fourthweb.ai)