And this is a possible reason to periodically change one's passwords on high-priority sites. Let's say my bank's servers get pwned and the attackers steal password hashes from the bank's servers. If it'll take them two months to crack my encrypted password from its hash, but I change it every month as a policy, then I've defeated the attack a month in advance without even realizing it.
From the previous article
Chiefcrowe posted, Hashcat also prioritizes on certain human-generated patterns, and millions of common human-generated passwords are already known by their hash based on the cracking of large batches of passwords in the past. Typical "leetspeak" substituting numerals for characters, or symmetrical or keyboard-patterned passwords, and so forth have been arbitrarily identified as easy prey.
Using an "inhuman" password like
KQ63m7pP2Jjw1$Q means they really will have to brute-force the whole keyspace to guarantee a solution, whereas
D3nv3rBr0nc0s is likely to be already known by its hash, based on cracking of previous batches of leaked passwords. Adding high-ANSI characters like ± and ™ force the attacker to take on an expanded keyspace as well. These techniques on an adequate-length password will make an attacker's job difficult.
The question is how to switch to strong "inhuman" passwords without losing your mind

Personally I use biometrics, namely a fingerprint reader with software that remembers my crazy passwords and auto-enters them in most situations with a finger swipe. I've heard good remarks about LastPass and KeePass too, but haven't tried them yet.