@charliermarsh: For fun, I also have 5.6 Sol building a CPython-compatible bytecode compiler. So far it's >40% faster with byte-for-byt…
Summary
Charlie Marsh announces he is building a CPython-compatible bytecode compiler that is over 40% faster with byte-for-byte compatibility on the majority of a large corpus, using Codex (AI) to autonomously optimize further.
View Cached Full Text
Cached at: 07/10/26, 04:13 PM
For fun, I also have 5.6 Sol building a CPython-compatible bytecode compiler.
So far it’s >40% faster with byte-for-byte compatibility on 64,535 out of 73,377 files in the corpus.
Codex is just iterating on it autonomously to further optimize and reach 100% compatibility. https://t.co/18jfcrrGXl
Similar Articles
@reach_vb: Codex for autonomously optimising your work:
Charlie Marsh shares that OpenAI's Codex autonomously found a 5-10% speed-up in a formatter while he was working on other tasks.
@charliermarsh: Optimizing Codex over the weekend. This is my favorite kind of work. I could do it forever.
Charlie Marsh tweeted about enjoying optimizing Codex over the weekend, sharing a link to the work.
@charliermarsh: Since my last post, I reduced ty’s retained memory by another 15% with Codex. We're now at a ~25% memory reduction over…
Developer @charliermarsh reports reducing ty's retained memory by another 15% using Codex, achieving a total ~25% memory reduction via /goal, and expresses enthusiasm for GPT models.
@charliermarsh: /goal for finding silly one-line optimizations that speed up your parser by 20-30%
Charlie Marsh shares a personal goal of finding simple one-line optimizations that can speed up a parser by 20-30%.
I built a compiler that rewrites Python into a model-facing representation
Vulpine is a compiler that transforms human-readable Python code into a compressed macro representation optimized for LLMs, reducing token count by 13.8% on average while enabling exact structural reconstruction.