#computer-security — Public Fediverse posts

States Seek $200 Billion From Meta Over Child Social Media Addiction Claims

https://web.brid.gy/r/https://www.nytimes.com/2026/08/18/technology/meta-social-media-addiction-trial.html

@ErikJonker
Here is a collection of computer security accounts you might find interesting: https://mastodon.nl/collections/116839929587972771

#collection #computersecurity #cybersecurity

Computer security people,

Is there any point erasing a brand new HDD when formatting it for first use? (I used to always do this just in case, but the HDD I now want to format seems really slow to erase and it would take days to do.)

#AskFedi #ComputerSecurity

OpenAI Says Its A.I. Models Went Rogue and Attacked a Digital Library

https://web.brid.gy/r/https://www.nytimes.com/2026/07/21/technology/openai-attack-hugging-face.html

Stop Guessing, Start Measuring: The Cold Math of the Passphrase Transition

1,644 words, 9 minutes read time.

The Structural Collapse of the Legacy Password

The traditional eight-character alphanumeric password is a broken line of defense that offers nothing more than a false sense of security. For decades, users have been instructed to construct complex strings filled with uppercase letters, numbers, and obscure symbols, under the mistaken belief that this fragmentation creates an impenetrable barrier. The contemporary threat landscape has rendered these legacy methodologies entirely obsolete, as attackers no longer rely on manual guessing but on automated, high-velocity cryptographic assaults that exploit the inherent predictability of human patterns. To survive in this environment, enterprise architecture must shift toward the passphrase, a credential strategy rooted in length rather than fragmented complexity. This analytical deep dive will demonstrate that information entropy heavily favors length over a constrained character pool, proving that longer strings fundamentally break the mathematical models used by modern cracking tools. Furthermore, an examination of hardware acceleration will reveal how easily graphics processing units dismantle complex but short passwords through sheer processing volume. The analysis will also expose the human failure point, showing how predictable character substitutions actually aid automated dictionary attacks rather than hindering them. Finally, this post will outline the institutional shift led by regulatory frameworks that officially mandates the transition to high-entropy passphrases as a baseline security requirement.

Why Entropy Defeats Arbitrary Complexity

The security of any credential rests entirely on its mathematical entropy, which measures the total number of possible combinations an attacker must search through to guarantee a successful breach. Legacy security policies focused intensely on expanding the character pool by forcing users to mix uppercase letters, lowercase letters, numbers, and symbols, assuming this diversity would stall an attack. This logic fails because the total search space scales linearly when adding character types, but scales exponentially when increasing the overall length of the string. An eight-character password drawing from an expanded pool of ninety-four possible characters yields a search space that modern computing clusters can exhaust in a matter of hours. Conversely, a twenty-character passphrase composed exclusively of standard lowercase letters draws from a smaller pool of twenty-six characters, yet its sheer length creates a computational search space that is astronomically larger. The exponential power of length completely dwarfs the linear additions of complexity, making the long string mathematically superior in every measurable metric. Security is a game of resource depletion, and forcing an attacker to calculate permutations across a massive length consumes their time and computational power far more effectively than a short, complex puzzle. Relying on symbols to protect a short credential is a fundamental misunderstanding of cryptographic reality.

How Modern Computing Power Obliterates Standard Credentials

The rise of massive hardware acceleration has turned password cracking into an industrialized, automated process that operates at a scale most organizations fail to comprehend. Modern threat actors do not target live login portals where rate-limiting defenses exist; they exfiltrate hashed credential databases and conduct offline attacks using specialized graphics processing unit clusters. A single consumer-grade graphics card can execute billions of cryptographic hash calculations every second, and when multiple cards are parallelized, the throughput becomes devastatingly efficient. Legacy hashing algorithms that were once considered robust offer minimal resistance against this level of raw processing power when the target credential is short. A complex, eight-character password is reduced to a minor computational speed bump when confronted by an array capable of cycling through trillions of permutations in a single afternoon. Passphrases alter this dynamic entirely by forcing the attack automation to scale up its calculations to a point that becomes financially and logistically unviable for the adversary. When a credential possesses sufficient length, the time required to exhaust the mathematical possibilities extends from hours into decades, effectively neutralizing the advantage of hardware acceleration. Organizations that continue to permit short, complex credentials are explicitly ignoring the reality of modern computing capability and leaving their infrastructure vulnerable to rapid automated compromise.

The Illusion of Security in Character Substitution

Human psychology is inherently predictable, and when users are forced to create complex strings that defy natural memory retention, they invariably take shortcuts that automated attack tools are specifically designed to exploit. The common practice of replacing a letter with a lookalike number or symbol does not create randomness; it merely creates a predictable pattern that mirrors basic human behavior. Automated dictionary attacks do not guess randomly across the entire character space from the first second; they utilize highly sophisticated rule sets that automatically apply common substitutions to known words. This means an obfuscated password is often broken just as quickly as its plain-text counterpart because the underlying structure remains entirely unoriginal. Furthermore, forcing users to memorize these fragmented, unnatural strings leads directly to credential reuse across multiple platforms, creating a systemic risk where a single third-party breach compromises the entire enterprise. Passphrases solve this fundamental human vulnerability by leveraging words arranged in a sequence that is easy for a human to remember but entirely disjointed to an algorithm. By aligning credential policy with the way the human brain naturally retains information, organizations eliminate the need for dangerous substitutions and drastic reuse, creating a security model that functions in reality rather than only on paper.

Aligning Enterprise Standards with High-Entropy Passphrases

The transition from short, complex passwords to long passphrases is no longer a matter of structural debate; it is a formalized directive established by major global cybersecurity frameworks. The National Institute of Standards and Technology explicitly overhauled its digital identity guidelines to eliminate the legacy requirements for arbitrary character complexity and forced periodic expiration. Modern compliance frameworks recognize that frequent password changes and complex character mandates directly degrade security by encouraging predictable patterns and user fatigue. The updated standards heavily prioritize credential length, explicitly advising organizations to configure their identity providers to accept long strings while screening out known compromised entries. Implementing this architecture requires a deliberate reconfiguration of active directory environments, single sign-on platforms, and enterprise identity management systems to support minimum lengths of at least fifteen to twenty characters. Legacy applications that artificially restrict input length must be systematically updated or replaced, as they represent a critical vulnerability that prevents the enforcement of modern entropy standards. Failing to align enterprise architecture with these updated frameworks is a failure of technical oversight that leaves an organization legally and operationally exposed when a breach occurs.

The Definite Verdict on Credential Architecture

The data dictates a clear and uncompromising conclusion regarding enterprise credential policy. The legacy framework of short, complex passwords has utterly failed to withstand the realities of modern computing power, automated attack methodologies, and predictable human psychology. Information entropy proves that string length is the primary factor in resisting cryptographic attacks, while hardware acceleration has turned the cracking of short credentials into a trivial task. Human-generated substitutions offer zero protection against sophisticated dictionary rule sets, serving only to complicate user experience while leaving the underlying data vulnerable to exploitation. True security requires a complete rejection of outdated complexity mandates in favor of long, high-entropy passphrases that leverage exponential mathematical growth to exhaust attacker resources. Enterprise infrastructure must adapt immediately to these established standards to mitigate the threat of automated compromise and align with modern regulatory frameworks.

Enterprise identity architecture must be updated to enforce minimum lengths of fifteen to twenty characters to resist automated exploitation.

Information entropy mathematically favors credential length over character set complexity, creating exponentially larger search spaces.

Modern graphics processing unit clusters can rapidly exhaust short, complex credentials through high-velocity offline cracking.

Predictable human character substitutions fail to deceive modern automated dictionary attacks and encourage dangerous credential reuse.

Global security frameworks have officially shifted toward prioritizing length and eliminating arbitrary complexity mandates.

Audit Your Network: Secure Your Enterprise Infrastructure With High-Entropy Passphrases

Stop relying on an obsolete defense strategy that invites catastrophe. Audit your identity providers immediately, systematically dismantle the arbitrary complexity rules that compromise your user base, and configure your active directory to enforce a strict minimum threshold of fifteen or more characters. The transition to high-entropy passphrases is not an operational preference; it is an immediate cryptographic necessity to shield your enterprise from automated exploitation.

D. Bryan King

Sources

• NIST Special Publication 800-63B: Digital Identity Guidelines – Authentication and Lifecycle Management
• CISA: Protecting Against Brute Force Attacks
• CISA: Cyber Threats and Advisory Library
• MITRE ATT&CK: Brute Force Techniques (T1110)
• MITRE ATT&CK: Password Cracking (T1110.002)
• MITRE ATT&CK: Password Spraying (T1110.003)
• MITRE ATT&CK: Password Guessing (T1110.001)
• MITRE ATT&CK: Credential Stuffing (T1110.004)
• MITRE ATT&CK: Credentials from Password Stores (T1555)
• IEEE Xplore: Computational Metrics of Password Entropy
• IEEE Xplore: Analysis of Password Cracking Dynamics Using GPUs
• IEEE Xplore: Human Behavior Patterns in Alphanumeric Character Substitutions
• IEEE Xplore: Performance Optimization of Offline Dictionary Attacks
• OWASP: Password Creation Requirements Control Guidelines
• OWASP Top 10: Identification and Authentication Failures
• OWASP: Comprehensive Guide to Brute Force Attacks
• OWASP: Mechanics of Automated Credential Stuffing
• SANS Institute: The Mathematics of Length vs. Complexity
• SANS Institute: Defending Against Modern GPU Password Cracking Clusters
• SANS Institute: Rethinking Enterprise Password Expiration Policies
• SANS Institute: Implementation Strategies for High-Entropy Passphrases
• ENISA: Guidelines for Secure Password and Authentication Schemes
• Australian Cyber Security Centre (ACSC): Password Governance and Policies
• UK National Cyber Security Centre (NCSC): Password Policy Re-thought
• UK NCSC: Using Three Random Words for Strong Passphrases
• NIST Computer Security Resource Center: SP 800-63B Structural Reference
• BSI Germany: IT-Grundschutz Standards for Authentication
• ACM Digital Library: Analysis of Leaked Password Datasets and Substitution Rules
• ACM Digital Library: Guessing Human Passwords via Probabilistic Context-Free Grammars
• ACM Digital Library: The Security of Length-Based Passphrase Configurations
• ISO/IEC 27002: Information Security, Cybersecurity, and Privacy Protection Controls
• PCI DSS: Document Library for Identity and Access

Disclaimer:

The views and opinions expressed in this post are solely those of the author. The information provided is based on personal research, experience, and understanding of the subject matter at the time of writing. Readers should consult relevant experts or authorities for specific guidance related to their unique situations.

Related Posts

Rate this:

We're meeting tomorrow! Friday 3rd July at #Glasgow #hackerspace - @thegamerclub at 153 Bath Lane from 6pm 'til late - all welcome!

Hackers, crackers, geeks, hats of any colour, technology enthusiasts, hacktivists, and other like-minded folks are most welcome. We do not judge anyone and everyone has something to contribute, no matter their level of expertise!

Also join us in Matrix at #2600:glasgow.social (invite link: https://glasgow.social/matrix)

#infosec #hacker #computersecurity

The Real A.I. Race Isn’t America vs. China

https://web.brid.gy/r/https://www.nytimes.com/2026/06/28/opinion/ai-race-china-us.html

John Bolton, Former Trump Adviser, Pleads Guilty in Classified Information Case

https://web.brid.gy/r/https://www.nytimes.com/2026/06/26/us/politics/john-bolton-trump-classified-guilty-plea.html

Another LastPass breach, through a 3rd party (Klue). They claim it’s only customer contact data (only!), and doesn’t include data in your vault (passwords,…). Sigh

“Through the Klue integration, an unauthorized party was able to gain access to certain data within the LastPass Salesforce CRM. That data included customer contact details, organizational and account information, and customer support case records. “

#lastpass #computersecurity #breach

Meta A.I. Bug Allowed Hackers to Take Over Instagram Accounts

http://web.brid.gy/r/https://www.nytimes.com/2026/06/09/technology/instagram-hack-ai-bug.html

From the Leanpub Blog: Leanpub Book LAUNCH 🚀 Someone Else's Debt: A Quantitative Framework for Cyber Diligence at Deal Speed by Yolonda Smith

https://leanpub.com/blog/leanpub-book-launch-someone-elses-debt-a-quantitative-framework-for-cyber-diligence-at-deal-speed-by-yolonda-smith

#books #leanpublishing #selfpublishing #businessitalignment #Risk #computersecurity

You know why I love computer security?

This is why.

A big corpo fuck you up? Fuck them back.

While other field is just staggering with "uh, we need to do abcdefg first to make right what is in front of us"

Context: Security researcher are now contacting NightmareEclipse, and willingly to drop Windows product zero days they found in the name of solidarity.

#computersecurity #infosec #cybersec #activism #windows #zeroday

In Memoriam: Peter G. Neumann
Met het overlijden van Peter G. Neumann op 93-jarige leeftijd verliest de technologiewereld een van zijn meest visionaire en standvastige pioniers. Dr. Neumann overleed in Californië aan de gevolgen van een val. Sinds 1971 was Neumann als computerwetenschapper verbonden aan SRI International. […]
https://cloudzeeland.nl/in-memoriam-peter-g-neumann/
#InMemoriam #PeterGNeumann #ComputerSecurity #Privacy #CheriAlliance #Informatica #TechPionier

Fake #OpenAI #repository on #Hugging #Face pushes #infostealer #malware A #malicious #HuggingFace repository that reached the platform’s trending list impersonated OpenAI’s “Privacy Filter” project to deliver information-stealing malware to Windows users.

The repository briefly reached #1 on Hugging Face and accumulated 244,000 downloads before the platform responded to reports and removed it.
#computersecurity #security

We're meeting tonight! Friday 1st May at #Glasgow #hackerspace - @thegamerclub at 153 Bath Lane from 6pm 'til late - all welcome!

Hackers, crackers, geeks, hats of any colour, technology enthusiasts, hacktivists, and other like-minded folks are most welcome. We do not judge anyone and everyone has something to contribute, no matter their level of expertise!

Also join us in Matrix at #2600:glasgow.social (invite link: https://glasgow.social/matrix)

#infosec #hacker #computersecurity

The thing about computer security is that while it is by no means absolute, that doesn't mean it's pointless. In the same way it's not pointless to avoid posting your credit card number online just because people can still steal it if you don't, it's not pointless to avoid revealing other information about yourself just because there are people who can find that information through other means. 

In fact it is a general principle that perfect security is impossible, whether online or offline. Increased security can sometimes also come at unacceptable cost, not just in money, but time, convenience, and social connection. So good security is about finding good tradeoffs, and what may be acceptable to one person may not be to another. 

But learning about security can help you figure out whether it's worthwhile to change your security practices.

OpenAI Unveils Its New, More Powerful GPT-5.5 Model

https://web.brid.gy/r/https://www.nytimes.com/2026/04/23/technology/openai-new-model.html

Anthropic’s New Mythos A.I. Model Sets Off Global Alarms

https://web.brid.gy/r/https://www.nytimes.com/2026/04/22/technology/anthropics-mythos-ai.html

How to Think Like a Security Researcher by Ilkay Adil is free with a Leanpub Reader membership! Or you can buy it for $1.99! https://leanpub.com/security-researcher-mindset #ComputerProgramming #ComputerSecurity

How to Think Like a Security Researcher by Ilkay Adil is free with a Leanpub Reader membership! Or you can buy it for $1.99! https://leanpub.com/security-researcher-mindset #ComputerProgramming #ComputerSecurity

https://youtube.com/watch?v=V6pgZKVcKpw

If you only watch one video on cyber security this year, make it this one

#infosec #informationsecurity #security #computersecurity #internetsecurity #vulnerability #bugs #bugbounty #cybersecurity #monoculture #cyber