Joshua Grochow on Nostr: Even in the noncommutative case, it is not known that 23 is the best you can do, it's ...

Even in the noncommutative case, it is not known that 23 is the best you can do, it's just that 23 is the best known to date! I believe the best known lower bound is still only 19 (due to Bläser; though I vaguely recall this was improved to 20 at some point but I can't find that at the moment.)