Claude Science

I want to evaluate and, if it fits, integrate a technique from recent AI research into my project.

## Technique: Claude Science
Claude breaks physics world record by solving nine-loop particle calculation unsupervised

Claude Science harness: SymPy + Python, 96 CPUs, 1 week, $1,000–$2,000. Solved the nine-loop scattering amplitude — previous record was eight loops, set by SLAC. One prompt, mostly unsupervised. Verified by Lance Dixon (the prior record holder). Demonstrates that frontier math/physics problems stuck for years can now be tackled at research-team budgets. Pattern is similar to Tab 40 (950 agents for biology) but focuses on exact mathematical symbolic computation rather than biological database search.

What the reference demo shows: 5-sub-tab Gradio demo — (1) Architecture: Claude Science harness stages (formulate → SymPy generate → verify → iterate); (2) Scattering Amplitude Context: loop expansion explanation and complexity growth table; (3) Mini Harness Demo: interactive 3-agent symbolic computation simulation; (4) Cost & Scale: loop-count vs compute/cost curve; (5) Implementation Patterns: 3 patterns for long-running symbolic computation agents.

## Reference implementation (Vector Ridge Labs, POC-043)
- Live demo: https://osoman2-gradio-pocs.hf.space/?poc=POC-043
- Source code: https://huggingface.co/spaces/osoman2/Gradio_pocs/tree/main/claude_science_harness
Original source: AlphaSignal Top News · 2026-09-28 · 6,823 likes

## My project
<Describe your stack, the problem you want to solve and where this technique would sit. Example: "Python FastAPI RAG service over 50k PDFs, Postgres + pgvector, OpenAI embeddings, p95 latency budget 800 ms.">

## What I need from you
1. Read the reference source code and explain in at most 5 bullets how the technique works and what it would change compared to my current approach.
2. Decide whether it fits my case. If it does not, say why and stop there.
3. Propose the smallest viable integration: files or modules to touch, new dependencies, configuration.
4. Implement it behind a flag or as a separate function so it can be compared side by side with what I have today.
5. Add an evaluation script that measures task success rate, number of steps and token cost per task, before and after, on my own data.
6. List the risks (latency, cost, failure modes) and what to monitor once it is in production.

Constraints: keep my existing stack and conventions, avoid heavy dependencies unless the gain is clear, and reference the file in the demo source whenever you reuse its logic.