Your job isn’t disappearing overnight.
But changes are definitely coming.
An Anthropic engineer says AI agents are about to reshape most work done on computers, and he expects the transition to be “painful” for a lot of people. According to a Business Insider report, Boris Cherny argues that software that can operate a computer on your behalf will spread beyond coding into the broader knowledge economy.
AI agents are already nudging work from “doing the task” to “directing the task.” Instead of writing every line, clicking every button, or chasing every file, workers increasingly describe work as: define the goal, hand it to an agent, then review and refine.
Why this could feel painful
Cherny’s warning lands because it’s not coming from a distant critic. It’s coming from someone close to the tools people are trying to operationalize inside companies right now.
The uncomfortable part is that computer-based roles have long felt insulated. Factory floors automated. Call centers offshored. Spreadsheets survived. Agentic software challenges that mental map by targeting the “middle layer” of office work: stitching together apps, moving information, drafting outputs, and iterating quickly.
Still, the story isn’t a clean replacement narrative. Recent research and benchmarks suggest agent performance in real-world tasks can be far shakier than hype implies, even as the tools keep improving.
What workers can do now
The most practical takeaway is to treat agents like a new co-worker archetype: fast, tireless, and occasionally wrong in a way that’s hard to spot.That makes three skills more valuable, not less:
- Scoping: Turning fuzzy asks into clear specs and constraints
- Supervision: Breaking work into checkpoints that an agent can’t skip
- Verification: Validating outputs with domain knowledge and judgment
If you want a north star for what’s changing fastest, look at software work. Engineers increasingly say AI is already producing huge amounts of code, shifting humans toward review, architecture, and decision-making. Zooming out, the bigger labor-market dynamic is churning: disruption and creation arriving together, with transitions that can outpace retraining.
Cherny’s “painful” prediction isn’t destiny. But it is a loud, inside-the-lab reminder: if your job lives in a browser tab, the workflow is going to get rewritten.
Also read: Microsoft’s view of AI trends to watch in 2026 shows how workplace agents, security, and governance are likely to collide this year.