If I understand things correctly - we can use 12 mnemonic words to generate a private key. Having 2048 possible mnemonic words it means that there will be 2048^12 possible states. Private key is 256 bit number so it will be 2^256. I have queried wolfram alpha against difference between these two numbers and it seems that private key space is greater (https://www.wolframalpha.com/input/?i=2%5E256+-+2048%5E12) so how is that possible that it is safe to derive private key from mnemonic words?
who said so? of course using private keys is the way to go. mnemonic words are for users, for the pros - there is a private key– NulikAug 4, 2021 at 21:43
FYI: decrypt.co/32853/…– NulikAug 4, 2021 at 21:45
The BIP 0039 specification says that for 12 words you get 128 entropy bits and 4 bit checksum.
So with 12 words you get 128 bits "security" and with 24 words you will get 256 bits "security".