Im not really against them, but, particularly these two seems like such a stupid huge waste of resources
Amicable Numbers : Only real world application, sharing a number with a loved one . So much for science.
Moo Wrapper : They try to break the RC5-72 algorithm, which hasnt been used for a long time. ETA? about 160 years.
Then there are the number theory ones, while not useless, as some have important real world applications, and some can lead to new theories that lead to another theories that lead to real world applications, they are just so boring to the average public. NFS, Numberfields, Collazt sieve, Yafu.
Special mention to PrimeGrid. While PrimeGrid is far from useless, prime numbers are important in science and specially in computing, we dont really need such huge numbers to this day, and computing for these numbers when we start to need them would be much more efficient. That isnt taking in account the possibility of Quantum Computing breaking prime number based cryptography in the long run.
I believe that we either shouldnt whitelist them, or we should count them as a single project for the rewards.