RE: https://infosec.exchange/@0xabad1dea/117475066389746858 The post-mortem has finally confirmed my intuition that this was a classic “confuse first byte of array for whole array” data processing bug. A few people were weirdly hostile about the suggestion that it was probably the same way we’ve been corrupting numbers for 50 years and not the newest, least efficient way to corrupt numbers. Unfortunately, gaining a dreaded new Nemesis has not made all of software engineering’s boring old rogue’s gallery retire