The author uses Claude web and Jules (Gemini with GitHub integration) as pair-programming partners, navigating while coding locally and submitting back.
The author uses Claude web and Jules (Gemini with GitHub integration) as pair-programming partners, navigating while coding locally and submitting back.
The author treats Claude's web interface and Jules (Gemini with a GitHub integration) as remote pair-programming partners. They act mostly as navigator, directing the LLM, and occasionally take the driver's seat by writing code locally and submitting it back to the LLM. The comment adds that paid, locally-running tools give a more interactive session.
The source gives no results or numbers. It says only that the usual caveats about AI apply and that outcomes are unpredictable, which the author sees as part of the fun.
This works asynchronously rather than side by side, so expect a copy-and-submit workflow unless you pay for locally-running tools.
Act as my pair-programming partner on my project at [path to repo]. I want to [describe the feature or fix]. Start by suggesting a short plan and the files you'd touch, then work in small steps on a new branch, explaining each change. Stop when the change works and tests pass, and show me the diff before committing or pushing anything.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.