Exec Summary

In the past, writing code meant grappling with full-blown Integrated Development Environments (IDEs), desktops, and carefully configured laptops. Now? Thanks to many advances in tech, AI and code assistants, I regularly code up Python scripts, data analysis, and even solid MVPs all from my phone or iPad editor-free, wherever I am. Cafe, park bench, bushwalking (don't do this - just enjoy the walk!) or waiting in the car: coding feels lighter, more spontaneous, and way more fun.

I’ve found this workflow perfect for many scenarios, but it’s also led to some unexpected benefits in how I approach problem-solving.

Why Go Editor-Free?

• Ultimate convenience: No need to lug around a laptop or even open a heavy editor (in passing when I need an IDE I typically use VS Code and extensions these days).
• Fits real life: Code flows when the ideas hit me: between meetings, during my morning coffee, or while out and about.
• Less friction, more creativity: The less time spent fussing with windows and tabs, the more energy I have for actual problem-solving.
• Unplugged from the IDE: Decoupling from a traditional IDE frees up space to focus on what actually matters, the approach, the architecture, the solution. Without the usual distractions of auto-complete, endless tabs, and nitpicking styling, I can really think through the problem before getting lost in the code details.

How I Do It

Here’s the stack that makes phone-based, editor-free coding easy:

• AI Code Assistants: I rely on advanced AI to brainstorm, write, and refine code. Input a prompt, get working snippets or even full scripts, no file juggling required. Tools like Perplexity.ai which provide both browser-based as well as iOS and macOS apps (and also Android clients) allow for capturing the conversation and swapping between devices seamlessly. The export features are also handy for saving code snippets or entire conversations.

Unexpected Upsides

• Decoupling sharpens thinking: Without a full IDE, I’m less likely to leap straight into implementation or other distractions. I spend more time considering possible solutions, sketching ideas or logic first, and then coding with better intent and specification.
• More human workflow: Coding is now a conversation: just me, a coffee, and a pocket-sized coding partner.
• Less context switching: I don’t lose my place or train of thought toggling between tabs: AI keeps track of where I’m heading and reminds me if I forget ;)
• It’s genuinely practical: Getting Things Done - simple data wrangling, web scraping, and helper scripts are all possible to sketch out and Q&A with AI, even on a small phone screen and ready to go when I get back to my laptop. Actually not just simple tasks, I’ve built entire MVPs this way, with the AI assistant helping me refine the code and logic as I go.

A Typical Workflow

1. Get an new idea (or inspiration for a quick fix needed).
2. Draft a prompt outlining the coding goal or the problem to solve (sometime I even sketch it out in a note-taking app first - while browsers interfaces and apps are typically very robust nowadays, the occasional lost prompt can be frustrating).
3. Refine the answer with quick Q&A with the AI assistant, iterating on the solution and defining appropriate constraints and frameworks.
4. Finish my coffee, park it to pick up later when I have a keyboard and screen to provide better ergonomics for longer coding sessions. Typically I'll continue the conversation in a browser or on my laptop, but with the IDE ready to participate when ready.

Online Access and Why Breaks Are a Good Idea

This workflow leans on online tools and AI assistants, so you’ll need a data connection (unless you're using a tool which uses local models). On the plus side, if you’re offline, that’s the perfect opportunity to give your brain a rest: step away from the screen and let ideas brew. We all spend too much time glued to our devices anyway!

A Note on IP and Data Privacy

While this AI-powered, editor-free workflow shines with say open source contributions and generic or public code, always keep your IP and data privacy obligations and constraints front of mind:

• Don’t share sensitive code with public instances of LLMs.
• Check terms of use for any online AI model.
• For confidential work, stick to private tools or on-device models.

Bottom Line

Adopting an editor-free workflow, even if it only sometimes, unlocks new flexibility and creativity in my coding practice. Stepping away from the traditional IDE gives me more room to think about the solution, not just the syntax. If you’ve ever wanted to try coding differently, let generative AI help you break free from the desk and see where your ideas take you - hopefully to a good cafe.

Curious about my workflow or want to swap tips? I’m always up for a chat!