The human is no longer in the loop, they’re in the cockpit

By Peter Springett · 23 April 2026

What the hell just happened? In the past few weeks, the way that I think, work and live with my laptop has changed completely. Maybe the biggest jump in the past four decades.

I’m currently working in the Claude app. Spinning ideas, stress testing workflows. Coming up with ideas that might evolve into an article like this one. Two years ago I would have opened a Word doc with Chrome on a second screen for research. More recently Chrome took second place to Claude and Gemini for research, planning and feedback.

But the past month changes everything. Almost without noticing, Claude has become my first screen. Instead of tabbing back and forth between app, browser, and chatbot, most projects now start and end with Claude.

Why now? Since the turn of the year, AI models have evolved from enthusiastic interns to highly capable colleagues. Then OpenClaw (autonomous AI) blew the doors off, triggering Anthropic and others to release similar software agents into the wild. We’re seeing the convergence of multiple tools into single platforms, Google patching NotebookLM into Gemini a good example.

From Research Tool to Primary Workspace

Critically, Claude and others offer project-style folders where files and chats co-exist in a persistent memory. The number of times where Claude now says, “Do you remember when Mark mentioned X, maybe we should reference it here?” is growing. Critical information no longer gathers dust in long-forgotten folders. It’s just a nudge away thanks to a co-worker with a largely infallible memory.

There’s much more to this than LLMs replacing apps and platforms one step at a time. It represents a shift in the way that I think, that my ideas evolve. Staring at a blank spreadsheet is anathema to imagination. Finite menus impose rules and constrain creativity. Multiple browser tabs muddle the mind. This model of working inevitably dulls the thinking process.

Strip away this framework and I can let my professional imagination run wild. No more files and folders or worrying about file extensions. I can advance unburdened, safe in the knowledge that my LLM is capable of spitting out the work in a variety of formats or coding an application that can be accessed at the press of a button. And if I am struggling to get out of first gear, my first prompt of the day is often, “Let’s pick up where we ended yesterday.” Or I can share ideas that are far from fully formed and discover a concept that has real legs.

Above all my thinking is more robust. Far from outsourcing cognitive discipline to an LLM, I’m more intellectually resilient. I lose focus less and I’m far more likely to respond to LLM goals and feedback. But I still retain agency and final authority over outputs and publication.

The 1984 Moment for AI

I’m well aware that this won’t be for everyone right now. It represents a paradigm shift as profound as the first Apple Macintosh operating system in the 1980s. Back then, a computer was largely for people who could code or at the very least were at home with a terminal style interface. Word processing and spreadsheets, however revolutionary, had minority appeal.

But within a decade most people were happy using a mouse to point and click at icons, paragraphs and worksheets. AI is doing the same thing, but one level up. It's making complex cognitive work legible and accessible to people who couldn't previously structure or execute it alone.

In 1984, nobody knew what the GUI would become. They just started pointing and clicking, and somewhere along the way the world changed. I feel the same acceleration now, every morning, when I open my laptop and Claude is the first thing I reach for. The files, the folders, the templates, they're still there when I need them. But I'm not working inside them anymore. I'm working above them. The human isn’t so much in the loop as in the cockpit.