@sophieschmieg@infosec.exchange we should go non-hybrid ML-KEM* plus ML-DSA** to avoid overflowing the ~1400 byte MTU!! * maybe not 1024 ** maybe not above 44