Recently, someone asked me: Why does on-chain data always seem to "pause" for a moment, even though blocks are already out? Basically, it's not that the chain is slow; it's that the data pipeline you're using is queuing. Indexers/Subgraphs need to first pull down the new blocks, run through parsing, and then insert them into the database. When your front end queries their database, not the chain directly; if there's a reorganization or a backlog on their side, it can feel like "there's no data just a moment ago, but a few seconds later, everything is caught up." Plus, RPC rate limiting makes it even more annoying: public nodes can't handle the requests and give you 429 errors or timeouts. Wallets, browsers, and aggregators all crowd together, making the experience feel like being stuck at a subway turnstile.

Later, I realized that the fuss over NFT royalties is somewhat similar: creators want stable income, secondary markets want smoother liquidity, but any delay or rate limit in the "settlement/data/execution" layers makes the user experience feel worse. Anyway, now I always check which subgraph I'm querying, whether the RPC is self-hosted or public, before blaming the chain right away.