How to use Claude to improve a language runtime
The author reports numerous runtime additions, expanded tests, parallel execution under two seconds, and performance improvements while reviewing every line.
Added1125 use casesShowing 169–192
Preparing search…
The author reports numerous runtime additions, expanded tests, parallel execution under two seconds, and performance improvements while reviewing every line.
AddedThe author reports that Claude resolved output differences between the C and C++ versions, including restoring original behavior where apparent fixes broke byte equivalence.
AddedThe author reports a two-month rewrite with fewer bugs, strong performance and easier feature development, while describing difficult reviews and substantial overtime.
Added FableThe author reports that test_failure_53 passed after Deebo investigated GROUPTOP-related failure hypotheses; the fix remained under review.
Added DeeboThe author reports several recent fixes passed tests and compiled on the first change set, with additional documentation, test, and configuration edits.
Added QoderA solo developer built a homeschool app support loop: parent bug reports become GitHub issues, Codex opens fix PRs, the developer merges, and parents receive in-app replies.
AddedA completed Mac wine cellar app uses bottle photos placed on Google Drive. Grok Bot retrieves details, four ratings, and tasting notes, writes them to a spreadsheet, and the app displays the collection.
AddedGrok Bot helped the author plan a marriage proposal and create Operation Liberty, combining a bot-narrated cinematic about the planning with a playable pixel-art game of the girlfriend's first Kuwait trip.
AddedLena handled launch work for a virtual try-on site built with Grok Build, producing a 25-brand target list, pitches, and two try-on Reels showing one outfit on three body types.
AddedGrok Bot helped set up a sister's Shopify store and created a phone-based barcode script. As the author scanned items, the bot researched them, wrote descriptions, and added inventory, with hundreds listed live in hours.
AddedThe author used Grok Bot to build an indexed library of existing offers, forms, and playbooks, redesign a branded one-pager PDF, build a fillable client prework form, and make a call sheet.
AddedThe author bought an old Dell Wyse on Grok Bot's advice, installed Ubuntu Server, and granted SSH access. The bot then set up Tailscale and installed a repository they had been working on.
AddedThe author reports 40,000 generated Rust lines and estimates 90% progress after two weeks of nightly iteration, while explicitly expecting additional fixes before the result works.
AddedThe article reports a completed Rust runtime rewrite across over 135 releases, with workload-specific performance and memory improvements and a few dozen resulting regressions.
AddedThe author estimates about 25% of issue-to-Copilot PRs are mergeable as-is and about 50% after comments, while reporting setup compliance and manual verification limitations.
AddedThe author and a coworker report fixing over 200 failing integration tests for parity with the old code in one week, followed by another code review before launch.
AddedThe author reports rebuilding the app, adding mobile clients with feature parity, and reducing monthly costs from about $20 to $0. They also describe remaining code quality issues.
AddedThe author reports queryable meeting notes across months of conversations and increasingly accurate summaries requiring less intervention as the context document accumulates corrections.
AddedThe author reports a core systems-language prototype in two weeks, followed by context and supervision problems that led them back to manual coding.
AddedA compliance writer uses Codex and Markdown to turn manually selected sources into section drafts, with repeated editing and final team approval.
AddedWillison independently built Cursor’s agent-generated FastRender browser on macOS and found a working window with legible pages and visible rendering problems.
AddedCursor replaced a flat shared-file coordination design with planners, workers and a judge, then ran large coding experiments including a browser and framework migration.
AddedDevClaw uses OpenClaw to run issue-based DEV and QA loops for isolated Telegram projects, reusing sessions and selecting models by task.
AddedThe author used GitHub Copilot with Opus and Gemini to construct a security research harness and carry a library bug through upstream remediation.
Added