Press "Enter" to skip to content

Twenty-One Years of Password Warnings: What the Hell Has Society Learned?

I’m going to ask a question that I’ve asked in different ways over the years. After decades of cybersecurity warnings, educational campaigns, podcasts, articles, and recommendations, what the hell have we learned about passwords?

I’m not saying that nobody has learned anything. Plenty of people have changed their habits. Some use password managers, multifactor authentication, and passkeys. Security professionals have developed better ways to protect accounts, and many organizations have improved their authentication systems.

But when I look at society as a whole, I have to wonder how much of that education is actually getting through.

Why? Because every year, somebody publishes another list of the most commonly exposed passwords, and every year, I recognize far too many of them.

The order changes. A few new entries appear. Some disappear and return later. But the same predictable numbers, words, keyboard patterns, and variations keep showing up.

I’m … the fuck … done with pretending this is a new problem.

Let’s go back to 2005

On September 8, 2005, Steve Gibson and Leo Laporte discussed passwords in episode 4 of Security Now!, titled “Personal Password Policy.”

Yes, episode 4. The podcast had barely begun.

In the original Security Now! transcript, they discussed predictable passwords, dictionary attacks, memorability, password reuse, and approaches to managing credentials.

They also discussed an issue that remains relevant today: when people are expected to remember too many complicated passwords, they may choose something easier to remember instead.

The discussion included the idea of keeping passwords in a protected location rather than forcing people to memorize everything. Software password wallets were already part of the conversation.

Some of the specific password-generation approaches discussed in 2005 are no longer what I would recommend today. Authentication technology and security guidance have evolved. But the underlying problem was already understood: people often choose predictable credentials because they’re convenient.

That was September 2005.

We’re now in November 2026.

Twenty-one years later, we’re still talking about people using passwords such as 123456.

Are you kidding me?

From password policies to password managers

Security Now! did not stop with its 2005 discussion. In July 2010, episode 256 examined LastPass, a practical response to the difficulty of maintaining different, strong passwords for many accounts. In June 2011, episode 303 explored Password Haystacks and the distinction between password length and predictability. These discussions matter because they show that the security community was not simply repeating a warning: it was investigating tools and techniques intended to change user behavior.

If we were identifying these problems more than twenty years ago, why are we still struggling with the same basic mistakes?

There is an important historical caution here. A 2005 password recipe, a 2011 discussion of password search space, and current authentication guidance are not interchangeable recommendations. Attack methods, computing power, password-storage practices, and standards have changed. The value of these older episodes is that they document how long the underlying problems have been recognized.

Fifteen years of password lists

Beginning in 2011, SplashData published annual rankings of commonly exposed passwords. Later, NordPass began publishing its own annual research.

These studies haven’t all used identical datasets or methodologies, so comparing one year’s numerical rankings directly with another year’s rankings requires caution.

But we don’t need to pretend that the studies are identical to recognize the recurring patterns.

Among the passwords that have appeared repeatedly in historical rankings are:

  • 123456
  • password
  • 12345678
  • 123456789
  • 12345
  • qwerty
  • admin
  • Other predictable variations of numbers, familiar words, and keyboard sequences

In January 2016, Computerworld published a comparison of the worst passwords from the preceding five years.

That article was already documenting the persistence of weak password choices.

Think about that. A decade ago, researchers and journalists were comparing multiple years of weak-password lists. Today, we’re still doing it.

Apparently, publishing another list hasn’t been enough to make the problem disappear. How many more annual rankings do we need before acknowledging that another warning, by itself, isn’t enough?

What happened in 2026?

On November 17, 2026, Cybernews published its coverage of the latest NordPass password study.

The research examined exposed credentials from public breaches and dark web repositories, covering the period from September 2025 through September 2026.

The three most commonly exposed passwords globally were 123456, admin, and 12345678.

According to The Register’s November 16 report, NordPass found that 123456 had occupied the top position in seven of its eight annual studies.

Seven out of eight years.

You have to be fucking kidding me. We’ve had years of warnings, better tools, and countless educational campaigns, and that six-digit sequence is still sitting at the top?

Before I even published this article, I asked a group of people which password they thought topped the 2026 rankings. Preston guessed 123456. He was right. On one hand, that’s encouraging: he recognized the pattern. On the other hand, what does it say about our collective progress when somebody can correctly predict the world’s most common exposed password without seeing the latest study?

That’s not a one-time mistake. That’s a recurring pattern.

The 2026 United States rankings included 123456, password, admin, 12345678, Password, 123456789, 1234, user@123, 12345, and Password1.

Notice something about that list?

Changing the capitalization of password doesn’t suddenly make it secure. Adding a 1 doesn’t magically make a familiar word unpredictable. And choosing a slightly longer sequence of consecutive numbers doesn’t solve the problem either.

We’ve known this for years.

So why are we still seeing these choices?

What the rankings actually tell us

There is an important distinction here. These lists are generally assembled from exposed or compromised credential datasets. They aren’t a complete survey of every password currently protecting an account on the internet.

That means we shouldn’t claim that the rankings prove every internet user has poor security habits. They don’t.

They also don’t establish that the same individual people have been making the same mistakes for fifteen years.

But they do establish that remarkably predictable credentials continue to appear in large collections of exposed passwords.

That is troubling enough.

It is especially troubling when the security community has been discussing these risks for decades and when safer alternatives have become more widely available.

My own history with this subject

I haven’t been sitting quietly while all this has happened.

On October 31, 2015, my Technology Podcast episode 224 included LastPass as part of a discussion about accessible applications and security news.

During National Cybersecurity Awareness Month in October 2019, I covered password managers, including my article NCSAM: Let’s start with password managers. Rereading that article now, I notice that I wasn’t telling everyone to use the same product. I was encouraging people to investigate the available managers and choose one that worked for them. The goal was education, not forcing someone away from a familiar tool without explaining why.

That same month, I was already asking a broader question in NCSAM: Is training to stay safe not sinking in? That piece was about phishing awareness and how people learn, not specifically about password rankings. Drawing on PhishLabs’ discussion of training programs, I argued for manageable lessons, different ways of presenting material, and an approach that fits the people being taught. I even compared the challenge to my own braille coursework: information doesn’t automatically sink in just because someone has presented it. Seven years later, I’m still asking whether cybersecurity education is reaching people and changing behavior. The fact that I’m revisiting the question in 2026 is part of what makes these recurring password rankings so frustrating.

Later that month, I discussed security features people should consider when using password managers. In November 2019, I asked whether cybersecurity education should stop just because October was over.

It shouldn’t, and I’ve never believed that it should.

On September 29, 2021, The Security Box episode 63 discussed password trends and the psychology of passwords.

On September 28, 2022, episode 114 introduced NCSAM with a discussion of the most common passwords of that year. The source material examined not only numerical passwords but also predictable choices based on subjects such as sports, food, and other familiar categories.

On October 4, 2023, episode 163 returned to passwords, password sharing, and multifactor authentication. Supporting blog posts included NCSAM week 1: How about them passwords? and Is it a good idea to share passwords? Not so fast.

In the first of those articles, I looked back at password habits from the DOS and early Windows years, discussed the move toward longer credentials and password managers, and encouraged readers to enable multifactor authentication. Some of the specific character-mixing advice in that 2023 piece differs from today’s NIST guidance. That is precisely why cybersecurity education has to evolve rather than endlessly recycling yesterday’s rules.

The second article dealt with a different kind of password problem: sharing credentials among family members or for streaming services. That’s not the same as choosing 123456, but it illustrates another reason password discussions can’t stop at telling people to invent a stronger string of characters. Account rules, convenience, and everyday behavior all matter.

On January 31, 2024, The Security Box episode 177 returned to password security. My supporting articles included Some of the dumbest passwords to use are … and This seems contradictory, don’t use long passwords, use unique complex passwords.

That episode also discussed password reuse, credential stuffing, and security incidents involving organizations.

And on December 3, 2025, The Security Box episode 263 returned to another list of passwords people should stop using.

Here we are in November 2026, discussing yet another list.

Have I been in this industry too damn long?

Do you see why I’m frustrated?

What the historical rankings reveal

The annual lists should not be treated as a scientific poll of all active accounts. The underlying collections can contain repeated credentials, old exposures, different geographic mixes, and passwords from services with different security requirements. SplashData and NordPass also used different sources and methods. That makes a single, uninterrupted year-by-year league table misleading unless the original methodology for each year is examined.

Nevertheless, the historical comparison is striking. In the 2011 SplashData list, password and 123456 occupied the first two positions. By 2013, 123456 had moved ahead of password. Years later, NordPass reported that 123456 ranked first in six of its first seven editions and then seven of eight in its 2026 report. That is not proof that nobody has improved; it is evidence that predictable choices remain stubbornly prominent in the exposed data.

There is also a useful generational finding. NordPass’s 2025 research reported that simple numerical sequences appeared prominently across the age groups it examined. Growing up with digital technology, by itself, does not guarantee good password habits. See NordPass’s research and methodology and The Register’s coverage of the 2026 edition.

It isn’t just about individuals

We also need to discuss the responsibilities of organizations and the services that accept passwords.

If a service allows somebody to choose 123456 as a password, why is that acceptable?

Why should a system accept one of the most predictable passwords in the world when it could reject it and require the person to choose something better?

If your system still accepts 123456 as a password in 2026, you have to be fucking kidding me. At some point, the responsibility belongs to the people designing these systems, too.

Organizations have responsibilities, too. They can implement protections against commonly used and compromised passwords, provide password-manager compatibility, support multifactor authentication, and adopt phishing-resistant authentication methods.

They can also stop creating unnecessary barriers that encourage poor password habits.

The National Institute of Standards and Technology addressed many of these issues in its July 2025 Digital Identity Guidelines, SP 800-63B-4.

Among other requirements, the guidance requires checking proposed passwords against lists of commonly used, expected, or compromised values. It also prohibits arbitrary character-composition rules and routine password expiration without evidence of compromise.

For passwords used as the sole authentication factor, the guidance specifies a minimum of fifteen characters.

That is a very different approach from simply telling people to change their passwords every few months and add an exclamation point.

The emphasis should be on choosing strong, unique credentials, rejecting known weak choices, and designing systems that help people make safer decisions.

Why does the problem keep coming back?

There probably isn’t one simple explanation.

Some people choose passwords they can remember easily. Others reuse passwords because they have too many accounts. Some don’t understand the risks. Others understand the risks but assume that nothing will happen to them.

A friend of mine prefers passwords he can remember because he doesn’t fully understand password-manager technology. He told me about an app he had found without realizing that it used a trust-no-one (TNO) approach. He later forgot the password to that manager and was locked out of Facebook. I don’t know whether the app description failed to explain the model clearly or whether the explanation appeared during setup and wasn’t understood. But his experience illustrates a serious problem: a tool can offer strong security and still leave someone stranded if its recovery limitations aren’t understood.

Some organizations still have poor authentication practices, and some services make it unnecessarily difficult to use stronger credentials.

Accessibility matters, too. A security system that creates unnecessary barriers for people using assistive technology can make good security practices harder to follow.

And then there is convenience.

People are busy. They want to sign in, complete their work, and move on. Security can feel like an obstacle, particularly when an application makes authentication more complicated than it needs to be.

But convenience doesn’t make 123456 a good password.

And repeatedly exposing accounts to predictable credentials isn’t a reasonable long-term solution.

What organizations should actually do

NIST’s current guidance is more specific than simply telling users to make passwords complicated. Services should reject passwords found on lists of commonly used, expected, or compromised values; limit failed authentication attempts; and protect stored passwords with appropriate salted, costly hashing. They should also allow password managers, autofill, and paste operations instead of obstructing them.

For centrally verified passwords used as the only authentication factor, NIST requires a minimum of 15 characters. Passwords used as part of multifactor authentication may be shorter, but must be at least eight characters. The guidance says not to require arbitrary mixtures of character types or periodic changes without evidence of compromise. See NIST SP 800-63B, Passwords.

These requirements also raise an accessibility question: can someone using a screen reader, braille display, or other assistive technology complete the authentication process without unnecessary obstacles? Password fields, recovery flows, autofill, and multifactor prompts should be accessible. Security that people cannot use reliably is security that may encourage workarounds.

We have better tools now

Password managers can generate and store strong, unique passwords without requiring people to memorize every credential.

Multifactor authentication can provide an additional layer of protection, although not every form of MFA offers the same resistance to phishing.

Passkeys offer another approach by replacing traditional shared passwords with cryptographic authentication. When implemented properly, they can reduce exposure to password phishing and reuse.

None of these technologies is a magical solution to every security problem. Accounts can still be compromised through other attacks, and implementation quality matters.

But the tools available today are considerably more capable than those available when Security Now! discussed personal password policies in September 2005.

So why do the same elementary password patterns keep appearing?

Cybersecurity awareness can’t be a once-a-year exercise

I’ve covered National Cybersecurity Awareness Month for years, and I’ve repeatedly made the point that security education shouldn’t stop when October ends.

We can’t spend one month telling people about passwords, phishing, and account security, then assume that everybody will remember the lessons for the rest of the year.

National Cybersecurity Awareness Month should cover more than passwords. That includes credit card security, banking notifications and why they matter, what people should and shouldn’t do, the consequences of ignoring warnings, and practical protections such as locking a payment card when appropriate.

I’ve covered these subjects through my blog and podcasts over the years. Search my blog for NCSAM and you’ll find plenty of articles and episodes. If you find a podcast you’d like to hear, tell me its episode number and I’ll help you obtain it. Older episodes may not be in the RSS feed, and not everyone uses RSS.

Education needs to be ongoing, practical, understandable, and accessible.

But education alone isn’t enough. Organizations must also make secure choices easier and insecure choices harder.

If people keep choosing predictable passwords, we need to ask whether the education is reaching them, whether the authentication systems are helping them, and whether the alternatives are practical for their circumstances.

I’ve worked in technology for years, so I understand much of this terminology. But how do we reach the masses with this training? Television and radio news generally give cybersecurity incidents sustained attention when they’re big enough to affect large numbers of people, whether that’s a major airline disruption, a Snowflake-related breach, or another widespread incident. Their reports tend to explain what happened, provide a few details, and tell people what to do next. That’s useful, but it isn’t the same as sustained security education. Written coverage may remind readers to use strong passwords, yet the broadcast discussion is often brief and general. Meanwhile, specialized technology coverage may already have moved on. How do we reach the people who aren’t following that coverage in the first place?

If the people who need this information aren’t hearing it, understanding it, or acting on it, are awareness campaigns accomplishing what they’re supposed to accomplish?

Repeating the same advice without examining why the problem persists isn’t going to solve anything.

So what the hell have we learned?

Security Now! was discussing password problems in 2005.

Annual weak-password rankings have been documenting familiar patterns since at least 2011.

I’ve discussed password managers, authentication, password reuse, weak credentials, and security education across my Technology Podcast, The Security Box, and my blog for more than a decade.

We’ve developed better authentication technologies. We’ve learned more about credential attacks. We’ve improved security guidance, and many individuals and organizations have made meaningful progress.

Yet in 2026, we’re still seeing 123456, password, admin, and other predictable credentials in prominent positions on the lists.

I’m not claiming that nobody has learned anything. That would be ridiculous.

I’m asking whether society, collectively, has learned enough to stop making the same elementary mistakes. Based on these lists and research that sounded a little too familiar to me, apparently not enough.

I’ve also posed this question to listeners and readers of The Security Box and The Security News. I’m interested in seeing how people respond, because this isn’t just a technical problem. It’s a question about education, behavior, convenience, and responsibility.

After twenty-one years of documented warnings, fifteen years of annual password rankings, and countless articles, podcasts, educational campaigns, and recommendations, what the hell have we learned?

Apparently, jack shit.


Discover more from Jared's Technology podcast network

Subscribe to get the latest posts sent to your email.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.