content(alternatives): differentiate intros across same benchmark cluster

[Flotapponnier]

Rewrites the intro paragraph on /alternatives/ pages that share the
same primary benchmark, so each page reads with provider specific facts
instead of templated paraphrases. Addresses the live audit finding that
/alternatives/alchemy and /alternatives/quicknode shared roughly 65 percent
of body content verbatim.

Constraints respected: no em-dashes, no marketing slop, 80 to 140 words
per intro, only verifiable facts. Auto generated leaderboard and shared
methodology footer below the intro unchanged.

Files touched: 8 of 22 alternatives YAMLs (the duplicate-clustered subset).

Clusters rewritten:

• aggregator-head-lag (3): alchemy, quicknode, pump-portal
• network-coverage (3): bitquery, coingecko, the-graph
• bridge-quote-latency (2): lifi, relay

The other 14 alternatives YAMLs either had already-differentiated long
intros (drpc, publicnode, tenderly, blocknative, etherscan, chainlink,
pyth, helius, jito, dydx, gmx, hyperliquid, jupiter, kalshi, polymarket)
or sit on a benchmark with no sibling page (birdeye on metadata-coverage,
dune on wallet-labels-coverage) so they cannot cluster as duplicates by
definition.