10-year dormant bug found in $XRP Ledger code.

The exploit required setting up hundreds of offers at absurd price levels, then triggering a single payment to consume them all. Normal payment flows never hit that code path — that's why it sat undetected for a decade.

Discoverer claims AI flagged it. Official response: no active exploit observed yet.

Question is — will AI start outpacing human auditors in finding edge-case bugs like this, or was this just lucky pattern matching?