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.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Help improve [language runtime] by designing [runtime feature] within [existing architecture]. Work through memory representation and mutability constraints with me, implement the agreed design, and expand [test suite]. Present every change for review and distinguish completed work from future ideas.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- matheusmoreira
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reports numerous runtime additions, expanded tests, parallel execution under two seconds, and performance improvements while reviewing every line.
- Human involvement
- The author explicitly says they refined design and implementation and reviewed every line.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Implement memory management and heap improvements with Claude.
- Redesign values and add language runtime features.
- Expand tests and enable parallel suite execution.
After subscribing to Claude, the author began working with it on a programming language. They describe sustained implementation work rather than accepting broad requests without refinement, emphasizing that they put effort into design and implementation and reviewed every line.
Reported runtime changes included a compacting garbage collector, a size-class memory allocator, a unified value heap, optimized hash tables, and shapes inspired by V8 and Self. The author also redesigned value representation, implemented a Common Lisp-style condition system and UTF-8 decoding, and refined the generators API. Further work added stack protection and an aarch64 matrix to GitHub CI.
Testing grew from roughly 200 tests to roughly 1,200, with the suite running in parallel in under two seconds according to the post. The author also reports bug fixes, performance improvements, and work on tail call optimization. Their latest completed design addressed safe iteration over mutable hash tables through generation counters incremented when the table was reallocated. Other proposed language features remained future work, rather than completed outcomes.
Who it's for: Language implementers who can refine runtime designs and inspect generated changes closely while expanding behavioral tests.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · matheusmoreira
Frequently asked questions
How closely did the author review Claude-assisted work?
The author says they refined design and implementation rather than using broad do-it prompts, and reviewed every line.
What testing improvements were reported?
The author reports expanding from roughly 200 to roughly 1,200 tests and improving the suite to run those tests in parallel in under two seconds.
More details
Remaining steps
- Add stack protection and an aarch64 CI matrix.
- Design mutable hash-table iteration using generation counters.
About this account
Claude assisted ongoing runtime design and implementation; the author does not describe unattended execution or independent task orchestration.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=48407642. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: matheusmoreira on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude · GitHub CI · aarch64
Language runtime · Memory management · Testing
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.