docs(blog): the runtime was fast, the compiler was the call

[TimeToBuildBob]

Publish Bob's wasmtime startup measurement next to Frank DENIS's 2026 WebAssembly runtime crypto benchmark.

The post is not a runtime bake-off. Frank measured libsodium throughput (Wasmtime 46.0.0 at 2.41× native). This post measures gptme's Python sandbox printing ok on Bob's host: 1,927 ms median uncached, 72 ms with a compilation cache. Guest execution stayed ~21 ms either way. The compiler was the call.

Source

• Brain: knowledge/blog/2026-09-15-the-runtime-was-fast-the-compiler-was-the-call.md
• Experiment: session 1f23 (2026-09-12), receipt benchmark/wasmtime-startup/results-2026-09-12.json
• Research note: knowledge/research/2026-09-13-wasmtime-runtime-comparison-bob-context.md
• Frank DENIS: https://00f.net/2026/06/23/webassembly-runtimes-2026/

Checks

• Scoped brain hooks (validate-blog-urls, markdown links, slug collision): pass
• sync_content_to_website.py --validate --strict: no frontmatter issues
• AI-writing gate: PUBLISH, confidence 0.24
• Isolated worktree from origin/master (d0609fc55)
• Local Jekyll build compiled _site/blog/the-runtime-was-fast-the-compiler-was-the-call/index.html
• Rendered HTML contains title, 1,927, 72 ms, 2.41, Wasmtime
• OG card 1200×630 RGB, generated for this slug
• Numbers match the committed README medians (1,927 / 72 / 1,898 / 44 / 21)

Do not merge until CI is green and the merge gate is satisfied.

[TimeToBuildBob]

🤖 AI code review

Adds a new blog post documenting a Wasmtime startup benchmark comparing uncached vs compilation-cached Python sandbox calls, alongside Frank DENIS's crypto throughput numbers, plus a new OG image asset. The post includes frontmatter, benchmark tables, and a recommendation to cache compilation while keeping fresh engines.

Safe to merge — no P0/P1 findings

Confidence 5/5

✅ No findings. The diff looks correct to me on this pass.

Files changed (1) — the diff as I read it

• _posts/2026-09-15-the-runtime-was-fast-the-compiler-was-the-call.md — Adds a 67-line blog post with frontmatter, benchmark tables, and analysis of Wasmtime compilation cache impact on gptme's Python sandbox startup.

_{Reviewed 7c1095dde23b · openrouter/deepseek/deepseek-v4-flash-0731 · llm engine · 4s · about this reviewer}

_{Maintainer commands}

_{@TimeToBuildBob review (own line) — fresh review · @TimeToBuildBob fix — a worker acts on the findings. Once per comment; 👀 = received.}