A top Anthropic engineer just said the quiet part out loud. Boris Cherny, a senior engineer working on Claude Code, told Business Insider that AI agents will transform every computer-based job in America — and it will be "painful."
This isn't another generic "AI will change everything" prediction. Cherny is building the tools that are already doing it. Claude Code, Anthropic's AI coding agent built on Opus 4.6, is one of the most capable agentic systems in production today. When the person building the weapon tells you it's dangerous, pay attention.
What Cherny Actually Said
The key points from the interview:
- Every computer-based job will be transformed — not just software development, but any knowledge work that involves a computer
- The transformation will be "painful" — this is not a gradual, comfortable transition. It's disruptive.
- Opus 4.6 represents a step-change — the latest Claude models have significantly improved agentic capabilities that make autonomous task completion practical
- Claude Code is the proof point — Anthropic's own coding agent demonstrates what autonomous AI agents can do when given the right tools
Why This Matters More Than Previous AI Warnings
We've heard "AI will transform jobs" a thousand times. Here's why this one hits different:
1. It's From an Insider, Not an Observer
Cherny isn't a pundit or analyst. He's a senior engineer actively building Claude Code. He sees the capability curve from the inside. When he says the transformation will be painful, he's basing that on what he knows is coming, not what he speculates might happen.
2. The Agentic Shift Is Already Happening
February 2026 has seen an explosion of AI agent tools and platforms:
- Claude Code Security — autonomous vulnerability hunting that finds bugs humans miss
- Microsoft Agent Framework RC — AutoGen + Semantic Kernel merged into a unified enterprise agent platform
- Samsung integrating Perplexity as a multi-agent system on Galaxy devices — AI agents going mainstream mobile
- NIST launching an AI Agent Standards Initiative — the government is standardizing this, which means it's real
3. The Funding Backs Up the Claim
In February 2026 alone:
- OpenAI: $100B+ funding round — the largest in tech history
- Anthropic: $30B Series G at $380B valuation
- Temporal: $300M for AI agent reliability infrastructure
- World Labs: $1B for spatial AI
That's $130B+ flowing into AI in a single month. The smart money isn't debating whether AI agents will transform work. They're betting everything on when.
What Actually Gets Transformed (And How)
Based on the current trajectory of AI agent capabilities, here's what's already changing:
Software Development: The First Domino
Coding agents like Claude Code, Cursor, GitHub Copilot, and Windsurf are already transforming how software gets built. Junior developer tasks are increasingly automatable. But it goes further — Claude Code Security can now autonomously hunt for security vulnerabilities, a task that previously required expensive specialized human expertise.
🔍 Explore AI Coding Agents
Compare 50+ AI coding agents in our AI Coding Agents directory — from Claude Code to Cursor to open-source alternatives.
Customer Service: Already Happening
AI agents handling customer inquiries isn't new, but the quality bar has shifted dramatically. Tools like Salesforce Agentforce and Intercom Fin now handle complex, multi-turn conversations that would have required senior support staff.
Sales & Marketing: The Next Wave
AI agents are moving from "write me an email" to "run my entire outbound campaign." Tools like 11x.ai and Clay are building autonomous sales agents that prospect, qualify, and engage — end to end.
Security: The Most Surprising
Claude Code Security finding vulnerabilities humans miss is a preview of a pattern that will repeat across every knowledge domain. When an AI agent can do expert-level security analysis autonomously, the value proposition of hiring a large security team changes fundamentally.
What You Should Actually Do About It
Panic is not a strategy. Here's what's actually useful:
- Start using AI agent tools now. The people who'll thrive in this transition are the ones who learn to work with AI agents, not against them. Browse our complete directory of 510+ AI agent tools.
- Focus on what agents can't do (yet). Strategic thinking, relationship building, creative direction, ethical judgment — these are still human strengths. Lean into them.
- Build on agent platforms. If you're a developer, the opportunity isn't just using agents — it's building on agent frameworks. Check out our guide to the best AI agent frameworks.
- Invest in understanding the ecosystem. The AI agent landscape is moving fast. Subscribe to our weekly newsletter to stay current.
📬 Stay Ahead of the AI Agent Revolution
Get the best new AI agent tools and industry analysis delivered to your inbox every week.
Subscribe Free →The Bottom Line
Boris Cherny isn't wrong. AI agents will transform every computer-based job. The question isn't whether — it's how fast, and how you position yourself.
The good news: the tools are available, many are free or low-cost, and the people who start learning now will have a massive advantage. The bad news: the window for "early adopter" advantage is closing fast.
Explore the full AI agent ecosystem: Browse 510+ AI Agent Tools →