I like how you’ve been saying for 3+ years that it’s possible to recover a SipHash key based on its output, something it’s specifically designed not to allow, but never given any evidence for it.