Tradeoffs

Because Bob cannot pre-generate addresses with silent payments, he needs to keep checking to find new payments from the point he generated the payment code. Because this scanning is relatively costly, Silent Payments require more compute and bandwidth when scanning than a standard Electrum-style server. The average Bitcoin wallet today will have to connect to a new type of server that serves the necessary “tweak” data for a wallet to check each potential transaction for themselves.

The key difference with Silent Payment scanning is that instead of pre-generating a large amount of addresses up front like with a standard BIP 32 light client, Silent Payments requires the wallet to download 33 bytes of data per potential output and then perform an ECDH calculation to check if it is owned by the user. The major benefit to this approach is that it provides excellent privacy (even for light wallets) as the wallet back-end does not know what outputs belong to any light client.

Even though this may sound like a major hit to user experience, thankfully we can already drastically improve sync performance by ruling out potential outputs like:

Non-Taproot outputs Taproot dust outputs <=1000sats (~85% of > Taproot outputs right now) All potential Silent Payments outputs spent since you last scanned Additionally, there are many brilliant people working on reducing the impact of this tradeoff through things like transaction cut-through, Silent Payments-specific indexes in Bitcoin Core, and much more.