Debug a phone app from a park bench
A developer spoke corrections into their phone while a coding agent began repairing a mobile web app outdoors.
- What they did
- While sitting at Albany Bulb at sunset, the developer checked the newest version of their mobile web app on a phone. They then used the phone’s voice input to tell their coding agent what was wrong, and the agent started drafting fixes right there.
- How it went
- They were able to keep working away from their desk, which they described as a revelation for working on the move. The source does not say whether the fixes solved the problem.
- Tips
- A phone with voice input and a data connection was enough for this setup; Starlink was mentioned only as a possible alternative. Build in a review step before any change is sent, deployed, or otherwise made final.
Try it yourself with Claude Code
I’m testing [project name] on my phone from [location]. Listen to my spoken feedback about what feels broken or awkward, then trace each problem to its source, make the smallest sensible fix, and verify it with the app’s tests or a build. After each change, pause and show me a short summary before you do anything irreversible such as sending, deploying, or deleting.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Deploy a project and register a domain from one agent flow
A custom agent can set up a Cloudflare account, buy a domain, and ship an app with only brief approval steps.
Add simple forgetting to agent memory
A developer added a basic forgetting layer to an agent’s memory and said it worked extremely well right away.
Shrink an AWS cost-triage MCP server
They rebuilt a read-only AWS cost checker with far fewer tools and cut the calls needed to answer a waste question from seven to two.
Auto-pick new bugs within team rules
A DeepSeek Harness agent watches one repo, follows team-written permissions, and opens work items only when the rules allow it.
Assemble a R coding agent from three basic tools
They wired a Claude agent in R to file and shell helpers, then hardened it with safer path checks and editing functions.
Convert a spoken brief into a working task tree
A developer fed a spoken project brief into LLM agents, which turned it into a spec, plan, scaffold, module tickets, and PRs.