This article was written with the help of AI, based on the author's original notes and ideas.

I recently told all my engineers to delete their IDEs. PyCharm, VS Code, Cursor. All of them.

Their replacement? Claude Code. A terminal. That's it.

The productivity spike

When I first made the switch myself, the change was immediate. Claude Code and later Claude Cowork felt like a superpower. Not a marginal improvement. A step function. I was shipping features in hours that used to take days.

I wasn't alone. Other founders and power users were reporting the same thing. You could see it in their commit histories: output that looked like their early days, the excitement was real.

The team's reaction

When I told the team, they were shocked. The immediate pushback was predictable:

"We'll lose control." How do we keep track of what's happening without a file tree and tabs?

"We'll have no overview." How do we navigate a codebase without an IDE's tooling?

"How do we even review the code?" If the agent writes everything, who checks it?

Fair questions. But I think they're the wrong questions.

The reframe

Why should we manually review every line an agent writes? That's the old way of thinking. Instead, build better tests. Build better CI. Build better acceptance criteria. Focus on outcomes and requirements, not line-by-line inspection.

If your tests are good, your linting is strict, and your requirements are clear, the code quality follows. The agent becomes reliable not because you watched it type, but because you built guardrails that catch problems before they ship.

So no, I haven't forced everyone to delete their IDEs yet. But I will. Especially as I'm now building Lanes with exactly this mindset.

"Coding is solved"

My gut feeling got confirmed when I heard Boris Cherny, Head of Claude Code at Anthropic, on Lenny's Podcast. He said he hadn't touched an IDE or written code by hand in months. His take: coding is practically solved for.

That was back in February. We're already seeing it play out.

The shift is already here

More and more engineers are working entirely from their terminals. They spin up agents, review pull requests occasionally, and spend most of their time thinking about what to build rather than how to build it.

But even this workflow is starting to show cracks.

The bottleneck is you

Here's what nobody talks about: when you run 5 to 10 agent terminals at once, you become the limiting factor. You lose track of what each agent is doing. Agents start fighting each other over overlapping files. Large contexts cause confusion. You're spending more time orchestrating than creating.

The agents aren't the problem. Your brain is.

That's the thought that led to Lanes. We need something to manage all this context, because we can't hold it all ourselves.

Why stop at code?

Take it one step further. When I connected all my tools through Claude Cowork, something clicked. Linear, Slack, docs, code, all accessible from one place. I felt invincible. No tab switching. No context loss. Everything in one flow.

So why can't the coding environment work the same way? Why are we still bouncing between five apps to do one job? Why does "development" mean juggling terminals, ticket trackers, chat, and documentation in separate windows?

It shouldn't. Not anymore.

Long live the ADE

The IDE was built for a world where humans type code. That world is ending.

What comes next is the agentic development environment: a single workspace where you think, delegate, and ship. Where your limited working memory isn't the ceiling on how much you can build. Where agents handle execution and you focus on direction.

The IDE is dead. Long live the ADE.