Anthropic's Top Engineer Says AI Agents Will Transform Every Computer Job

Boris Cherny, the lead engineer behind Claude Code at Anthropic, just said what many in tech have been thinking but few are saying out loud: AI agents will soon handle any kind of work you can do on a computer — and the transition is going to be "painful."

This isn't vague futurism. Cherny is the person who built the system that lets Claude operate computers like a human. He knows what's coming because he's building it.

What Cherny Actually Said

In a February 2026 interview on Lenny's Podcast, Cherny explained that AI agents are rapidly evolving beyond generating text or answering questions. They're now capable of executing tasks across workplace tools — email, spreadsheets, project management systems, databases, whatever you use.

According to Business Insider (February 22, 2026), Cherny said the job title "software engineer" will start to disappear in 2026. Not eventually. This year.

He continued: "In the meantime, it's going to be very disruptive. It's going to be painful for a lot of people."

This is the engineer who created the tool talking, not a pundit speculating from the sidelines.

What Claude Code Actually Does

Claude Code is Anthropic's AI coding agent built on top of the Claude Opus 4.6 model (released February 2026). Unlike chatbots that generate code snippets, Claude Code can:

• Navigate file systems and edit multiple files
• Run terminal commands and debug errors
• Search documentation and implement solutions
• Commit code to version control systems
• Interact with web browsers to test changes

It's not generating suggestions. It's doing the work — start to finish.

The key difference: previous AI coding tools were assistants. Claude Code is an agent. Assistants help you work faster. Agents do the work for you.

From Coding to Everything Else

Cherny's point is that what works for coding works for any computer-based job. If an AI can navigate file systems, run commands, and interact with applications, it can:

• Manage your inbox and calendar
• Update spreadsheets and databases
• Process invoices and expense reports
• Monitor systems and respond to alerts
• Research competitors and compile reports
• Schedule meetings and coordinate teams

The technology isn't waiting for permission. It's already here. The question is how fast businesses adopt it.

Why This Is Different Than Previous Automation Waves

We've had automation scares before. Spreadsheets replaced accounting clerks. Email reduced the need for secretaries. Cloud computing eliminated on-premise IT roles.

But those changes were narrow — specific tools replacing specific tasks. AI agents are general-purpose. They can learn new tools, adapt to new workflows, and handle ambiguous instructions.

A traditional automation script does exactly what you program it to do. An AI agent can figure out what needs to be done and do it — even if it's never seen that exact task before.

The Speed Factor

According to Benzinga (February 22, 2026), Cherny noted that these systems are "rapidly evolving." We're not talking about a 10-year transition. The technology is improving month by month.

Anthropic released Claude Opus 4.6 in February 2026 with significantly better computer use capabilities than the October 2025 version. That's massive progress in four months.

If the improvement curve continues, we're looking at enterprise-grade AI agents handling complex workflows by late 2026 or early 2027.

What "Painful" Actually Means

Let's be specific about what Cherny is warning about.

Job Displacement Will Be Real

Roles that are primarily computer-based — data entry, customer service, bookkeeping, basic programming, administrative coordination — will shrink. Not because AI is better at creativity or judgment, but because it's faster, cheaper, and doesn't need training.

A company paying $50,000/year for an admin assistant can replace 80% of that work with an AI agent that costs $500/month in API fees. The math is brutal.

Reskilling Won't Be Fast Enough

The standard response to automation is "people will learn new skills." That's true long-term, but short-term it's messy.

If you're a mid-career accountant whose job gets automated, you can't just become a machine learning engineer overnight. Retraining takes years. The transition will hurt.

The Gap Between Early Adopters and Everyone Else

Companies that deploy AI agents early will have a massive efficiency advantage. They'll produce more with fewer people, undercutting competitors who stick with traditional workflows.

This creates a forced adoption dynamic — you either use AI agents or get out-competed by companies that do.

What This Means for Business Owners

If you run a business, Cherny's comments aren't a threat — they're a roadmap.

Don't Wait for "Perfect" AI

AI agents don't need to be 100% accurate to be useful. They need to be 80% accurate and 10x faster than humans. We're already there for many tasks.

If you're waiting for AI to be "ready," you're already behind.

Start With High-Volume, Low-Risk Tasks

The smart play is to deploy AI agents for repetitive work where mistakes are easy to catch:

• Email triage — sorting, flagging, drafting routine replies
• Data entry — extracting information from documents into systems
• Scheduling — coordinating meetings, booking calls
• Monitoring — watching dashboards and alerting on anomalies
• Research — gathering competitive data, market intel

These are tasks where AI saves hours every day, but human oversight catches errors before they cause damage.

Your Team Will Resist — Prepare for That

When you introduce AI agents, employees will worry about their jobs. That's rational.

The framing matters. Position AI as eliminating boring work, not eliminating jobs. The people who embrace working with AI agents will become more valuable. The ones who resist will get left behind.

The Jobs That Won't Be Automated (Yet)

Not everything is at risk. Here's what AI agents still struggle with:

High-Stakes Decision-Making

AI can analyze data and suggest options, but final calls on strategy, hiring, major investments — those still need human judgment. Trust and accountability matter.

Complex Negotiation

Closing deals, resolving conflicts, managing sensitive relationships — these require empathy, reading between the lines, and adapting to emotional dynamics. AI isn't there yet.

Creative Strategy

AI can execute creative briefs, but defining the vision, understanding cultural context, taking bold creative risks — that's still human territory.

Physical Presence Work

If your job requires being in a specific place — construction, healthcare, hospitality — you're safe from this wave of automation. AI agents live in computers, not the physical world.

How We're Approaching This at SetMyClaw

We help businesses set up AI assistants — specifically using tools like Claude and OpenClaw. When we talk to clients, the conversation always starts the same way:

"What are you currently paying someone to do that a computer could handle?"

The answer is usually 30-50% of administrative work. Email management. Calendar coordination. Data cleanup. Report generation. Monitoring systems.

Our approach:

• Start small — One workflow, not a full replacement
• Measure impact — Hours saved, errors caught, ROI
• Iterate — Expand what works, drop what doesn't
• Keep humans in the loop — AI does 80%, humans verify the 20% that matters

The goal isn't to fire your team. It's to make your team 3x more productive.

The Timeline: When This Actually Happens

Cherny said software engineering roles will start disappearing in 2026. That doesn't mean every developer loses their job this year. It means the shift has started.

Here's the likely progression:

2026: Early Adopters

Tech companies, startups, forward-thinking enterprises start deploying AI agents for internal ops. Junior roles shrink. Productivity per employee jumps.

2027: Mainstream Pressure

Companies that ignored AI agents see competitors moving faster and cheaper. Forced adoption begins. Mid-level roles get restructured around AI oversight.

2028-2030: New Equilibrium

AI agents are standard. "Computer work" no longer means humans sitting at keyboards — it means humans directing AI systems that do the actual execution.

The people still employed are the ones who learned how to manage AI agents, not the ones who tried to compete with them.

Bottom Line

Boris Cherny isn't predicting the future. He's building it. And he's telling you it's coming fast.

AI agents will transform any job you can do on a computer. The transition will be disruptive. Some people will lose jobs. Some companies will thrive. Some will get left behind.

If you're a business owner, the question isn't "Should I use AI agents?"It's "How soon can I start, and where do I deploy them first?"

The painful part isn't the technology. It's watching your competitors adopt it while you wait for the "right time."

The right time is now.