I should not be allowed into the same room as crypto development, and have certainly never tried to "attack" a crypto algorithm. Still, reading that, against 2

I should not be allowed into the same room as crypto development, and have certainly never tried to "attack" a crypto algorithm. Still, reading that, against 22.5 (of 24) rounds of the Gimli computation, the attack is claimed to need 2^129 bits of memory. That is 77,371,252,455,336,267,181,195,264 TB if math is right, which does seem to gently push that "attack" into a rather theoretical plane? Not sure what I'm missing, from your tone I would expect a smoking hole and this doesn't seem to be that.