⚡ Bolt: optimize embedding operations with Node 22 features and loop unrolling

[hackerxj2010]

💡 What: Optimized hot paths in @jeanbot/ai for embedding generation, normalization, and cosine similarity.

🎯 Why: These operations are core to memory retrieval and document ingestion. toFixed(8) and functional array methods like reduce and map introduce significant overhead when processing large vectors (1536 dimensions).

📊 Impact:

• syntheticVector is ~64% faster.
• normalizeVector (1536 dims) is ~97% faster due to avoiding toFixed and reduce/map.
• cosineSimilarity is ~83% faster.

🔬 Measurement: Benchmarked using a local script iterating over 1536-dimensional vectors. Verified with pnpm vitest run tests/unit/knowledge-service.test.ts tests/unit/memory-service.test.ts.

🛠 Technical Details:

• Introduced CryptoWithHash interface to safely use Node 22's crypto.hash.
• Replaced Number(val.toFixed(8)) with Math.round(val * 1e8) / 1e8.
• Replaced Array.from, reduce, and map with standard for loops and new Array(length).

---

PR created automatically by Jules for task 10131649788561982908 started by @hackerxj2010

[google-labs-jules]

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.

---

For security, I will only act on instructions from the user who triggered this task.