Post #3665895

I was rejected by the International Association for Cryptologic Research. They said submissions should ‘be clear, readable, and self-contained’, ‘look somewhat new and interesting’, and ‘contain proofs or convincing arguments for any claims’. I was none of these. The academic establishment is still blindly trusting hardware-backed secure boot, treating it like an infallible standard wrapped in formal proofs. But those proofs are useless in the real world because secure boot relies on logic branches—and logic locks can be glitched. We have proven that physical fault injection shatters these systems. A $5 voltage glitcher can drop the core voltage for a fraction of a microsecond, force the silicon to skip the signature verification instruction, and bypass the entire security model. The hardware just forgets to check the lock, and the execution door swings wide open. I tried to explain that the 'proof' against this isn't some elegant equation—it's the fact that the execution is the decryption. If an attacker tries to fault-inject the Speck cipher, it doesn't skip a vulnerable logic gate; it spits out garbage plaintext, fetches invalid opcodes, and hard-faults the processor. The entire point is that the machine code is inherently unreadable without the correct cryptographic state. There is no master key. The payload is the ciphertext. But apparently, actually building a bare-metal framework that turns physical exploits into a mathematically unsolvable problem isn't 'academic' enough unless it's wrapped in 15 pages of theoretical fluff.