docs(readme): refresh Why/How-It-Works story and trim visual noise

- Why: open on the pain, then a real before/after (raw requests+XBRL
  vs two lines) instead of the abstract "turns filings into objects"
- How It Works: add the Company/Filing/.obj() mental model in words,
  plus a pipeline/architect beat (warehouse, MCP, scale)
- Quick Start: fix paste bug (pip install was inside the python block)
- Use Cases: lead with Financial Statements; reorder 13F to third
- Re-point feature cells at the pipeline audience (Pipeline Ready, RAG)
- Trim hexagon dividers 9 -> 3 (keep only narrative pivots); drop the
  unsubstantiated "10x Faster" badge
- Fix broken examples: form4.transactions and facts.to_pandas() do not
  exist; use form4.to_dataframe() and facts.query().by_concept(...)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>