Lately I’ve been tinkering with on-chain data and found that a few node RPCs return different results... For the same block, some say it’s finalized, while others still have it pending. To put it plainly, what we call “on-chain” is actually indexed and synced through systems, so there’s always some delay. Especially with those tag/label systems—you think it’s real-time, but they might only capture a snapshot every few minutes. So sometimes when you see an address transfer, it may have already happened a few minutes ago. Recently, several tools have been criticized for being laggy, and honestly that makes sense. Anyway, now I always compare multiple sources when looking at data, or even run a node myself. Has anyone tried running their own light client? It feels like it could help avoid a lot of “false impressions.”