I’ve been meaning to write something about passwords, and password managers for ages.


  1. Passwords need to be as long as possible, and no fewer than 15 characters.
  2. Never re-use the same password on different sites.
  3. Always use 2FA where it’s offered.
  4. Use a password manager like bitwarden to do all the above!

There really is no excuse these days for having poor password hygiene, such as re-using passwords, using easily guessable or crack-able ones, not changing them periodically. A good password manager makes life incredibly easy – I no longer know what most of my passwords are, because bitwarden looks after them all. It integrates with my browser, and with my phone, and gives me excellent passwords and (with a paid plan) 2FA for all the sites I use.

I use a paid plan with Bitwarden to get the premium features like 2FA, analysis tools, and sharing passwords with multiple users. The free plan is also really good though, and I started off with that. At time of writing, the individual paid plan is $10/year, but the families plan (which I have) is only $40/year, which lets me and 5 other users join my ‘organisation’ and get password sharing and the premium features. They otherwise all have free Bitwarden accounts.


First of all, why do we need good passwords? Fundamentally we need good passwords because there are crackers in the world who want access to your account/data/money/… They may do this by guessing the password, or getting you to tell it to them, or by stealing the login database/buying it off the dark web. If we have decent passwords, the first two should be avoidable – but sooner or later a database with your password in will be compromised.

Thankfully the days of storing passwords in plain text in the database are long gone – after all it doesn’t matter how good your password is if it can just be read from the database! Instead, passwords are usually stored as a hash, which is a one-way encoding of the data. That it is say, it is impossible to work backwards from the stored hash to the original password. When you enter the password the first time, it is hashed and stored in the database. The next time you try to log in, you enter the password, which is hashed again, and the new hash is compared with the old. If it matches, you’re in!

As a very simple example, if you assign ‘a’ the value 1, ‘b’ the value 2, and so on, then a hash might be adding up all these values. So if the password is ‘aaab’, then the hash will be 5. It is impossible to start with the 5 and work back to ‘aaab’. Even if you know the password had 4 characters, you can’t tell if it’s ‘aaab’, ‘aaba’, ‘abaa’, or ‘baaa’. This strength is also a weakness, as any of these the passwords, as well as ‘baaa’, ‘bba’, ‘e’, ‘da’ all also hash to 5 – a so-called hash collision. This means if someone tried ‘e’, they would be able to log in, as it hashes to the same value.

In reality, hashes (or these days key-derivation function) are super clever and complex, so the chance of a collision is vanishingly low – and they are robust to analytical weaknesses (unlike our example above). We’re not completely out of the words though, because if you have the hash it becomes theoretically possible to crack passwords by brute-force. Trying out every combination of letters and seeing which one gives a hash match will reveal the password (or at least one which gives the same hash).

This is where password length and (to a lesser extent) complexity come into play. Modern computer hardware can generate hashes of all possible 8 or less character passwords in minutes, if the system is using an older hashing function. The same hardware will take decades to do the same for 15 character passwords.

In other words, waterhorselunch is a much stronger password than 7dYnV*fd. And also much easier to remember and type in, especially on a mobile device.

I would still say waterhorselunch has some analytical weaknesses – there are ‘only’ 200,000 words in English, so brute forcing all possible 3 words combos isn’t too hard. So spice it up with a separator, a number, and some capitals, and the password is good – Water|horse1|lunch. Still almost as easy to type in, but now not susceptible to a dictionary attack. Even better is 4, or 6 words.

Of course, if all always use waterhorselunch, or even Water|horse1|lunch, then as soon as it’s cracked once, the cracker can log in to all your accounts. So don’t do it!

Password Managers

It doesn’t matter how strong your password is if you either can’t remember it, or if you re-use it on different sites.

This is where passwords managers, like bitwarden, come in. Essentially they generate and store all your passwords for you. The only password you have to remember is the one to unlock the vault. Now when I log into LinkedIn, I click on my Bitwarden browser extension – LinkedIn is already shown, and a simple click populates my username and password.

Bitwarden will generate strong, long passwords, with a high level of customisation. You can also keep generating new ones until you find ones you like. And edit them further (so if a specific site needs a particular character, you can add it).

Here are some examples:

  • hatchback.unlovely5.dramatic
  • estrogen3.mushroom.satisfy

Alternatively, you generate ‘traditional’ passwords like WA2^^cbtooA5Vw2GdfMj7i2$roRxiPaupBmDUJq. But I know which I’ve rather have to type in using an PS5 controller!!

Bitwarden will also tell me where my details have been involved in a known data breach, and if I have re-used passwords, or am using passwords from a common list.

It’s worth mentioning that Bitwarden don’t have access to my passwords. They are encrypted with a very strong algorithm that is keyed on my master password, and it is only ever this encrypted data which is sent to/from their servers. The downside is that if I do ever forget my master password, or lose my YubiKey (see below), then I can’t access my passwords anymore. But equally that is kind of the point!

Two Factor Authentication

The other major weapon in the fight against crackers is multi-factor authentication (for our purposes two-factor authentication, or 2FA). The first factor is your password (something you know). The second factor is a code (generated by a device only you have). Multi-factor authentication is a whole subject in itself, but for most normal people it boils down to having to enter a time dependent code (usually 6 digits), either generated by an app (such as Google Authenticator), a piece of hardware (like the card readers banks used to use) or sent to you by SMS/e-mail. The latter option is a bit rubbish, as neither SMS nor e-mail are secure. 2FA is super secure because only the person with the device can generate the code, each can only be used once, and it changes every 30s so can’t really be stolen. Biometrics (fingerprint login) can also be a second factor.

The trouble with something like Google Authenticator though is that you need your phone (or whatever device it is on), and it’s a bit of a pain to be honest. Once or twice in the past I accidentally left my phone at home, and wasn’t able to log into the AWS Console at work as a result!

So I now use Bitwarden to store my authentication codes, which it does superbly.

This is slightly breaking the principle of 2FA, as there is arguably only one factor involved. If someone accesses my Bitwarden account, they have access to both factors (password and code). However, if my passwords are compromised outside of Bitwarden, then my accounts are still protected by the second factor. This is good enough for me, for the convenience of having 2FA protection without it being a pain for me.

I also have a YubiKey as 2FA on my Bitwarden account, so that even if someone does get hold of my master passphrase, they still can’t access my account. In that sense, it is kind of 2FA by proxy – you need the YubiKey, it’s just one step removed.