This exploit can't be disproved by saying "read the code". The point is that a CPU bug causes the code to run off the rails and do something that it isn't even

This exploit can't be disproved by saying "read the code". The point is that a CPU bug causes the code to run off the rails and do something that it isn't even written to do. Examining the source code shows an infinite loop running along. Dumb and harmless, right? Not if a CPU bug corrupts the L1 cache memory in which the loop is running. Now that loop could jump anywhere and do anything. You can never prove behavior at that level from abstracted high-level source code alone. As for the Hacker News reaction, whether the exploit actually works is somewhat immaterial. Even if it doesn't, the story opens up new and interesting lines of thinking around CPU architecture and security. It's interesting hacker material even if there's no bug here.