<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Cybersecurity on Ulveon's Thoughts</title><link>https://ulveon-thoughts-f210db.gitlab.io/category/cybersecurity/</link><description>Recent content in Cybersecurity on Ulveon's Thoughts</description><generator>Hugo</generator><language>en-IE</language><lastBuildDate>Tue, 05 May 2026 14:00:00 +0100</lastBuildDate><atom:link href="https://ulveon-thoughts-f210db.gitlab.io/category/cybersecurity/index.xml" rel="self" type="application/rss+xml"/><item><title>Why did AI destroy my production database?</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2026-05-05-why-did-ai-destroy-my-production-database/</link><pubDate>Tue, 05 May 2026 14:00:00 +0100</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2026-05-05-why-did-ai-destroy-my-production-database/</guid><description>&lt;p&gt;I already posted &lt;a href="https://ulveon-thoughts-f210db.gitlab.io/p/2025-08-10-ai-is-not-a-fad-its-here-to-stay/"&gt;my thoughts on AI&lt;/a&gt; and why I don&amp;rsquo;t think it&amp;rsquo;s going away any time soon. Unfortunately, it seems some people who don&amp;rsquo;t like LLMs are using AI-induced outages and deletions as an opportunity to reaffirm their biases, and, in doing so, may be missing part of the picture.&lt;/p&gt;</description></item><item><title>The secure open source fallacy</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2025-12-03-the-secure-open-source-fallacy/</link><pubDate>Wed, 03 Dec 2025 08:00:00 +0100</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2025-12-03-the-secure-open-source-fallacy/</guid><description>&lt;p&gt;Most open source advocates, and many security professionals, often say things like &amp;ldquo;open source software is secure because you can just read the code&amp;rdquo;.&lt;/p&gt;
&lt;p&gt;This argument assumes that the ability to read source code directly translates into the ability to understand, verify, and trust it, because you can see the files this software opens or the network sockets it listens on. You can see the kind of network data it sends, and the cryptography it uses.&lt;/p&gt;</description></item><item><title>Vanguard (Part 1): Telegram raid protection</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2025-08-09-vanguard-part-1-telegram-raid-protection/</link><pubDate>Sat, 09 Aug 2025 00:00:00 +0000</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2025-08-09-vanguard-part-1-telegram-raid-protection/</guid><description>&lt;p&gt;I&amp;rsquo;ve been developing Vanguard over the past few months. Initially it was little more than a hack, which was developed quickly and without much consideration regarding the cleanliness of the code. I needed a way to have public Telegram links without the risk of raids, or getting botted.&lt;/p&gt;</description></item><item><title>Effective security beyond encryption</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2025-05-25-effective-security-beyond-encryption/</link><pubDate>Sun, 25 May 2025 00:00:00 +0000</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2025-05-25-effective-security-beyond-encryption/</guid><description>&lt;p&gt;To highly technical audiences, it is unsurprising that Telegram is not cryptographically secure and that Signal is much better.&lt;/p&gt;
&lt;p&gt;Numerous security experts have pointed out that Telegram&amp;rsquo;s reputation as a secure messenger is primarily a result of its marketing rather than its technical architecture.&lt;/p&gt;</description></item><item><title>Software toggles are evil</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2025-05-13-software-toggles-are-evil/</link><pubDate>Tue, 13 May 2025 00:00:00 +0000</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2025-05-13-software-toggles-are-evil/</guid><description>&lt;p&gt;&amp;ldquo;Do not sell my personal information&amp;rdquo;, &amp;ldquo;Reject all cookies&amp;rdquo;, &amp;ldquo;Do not show me personalised ads&amp;rdquo;, &amp;ldquo;Do not use my Maps activity for improvements&amp;rdquo;, &amp;ldquo;Do not store my voice recordings&amp;rdquo;, &amp;ldquo;Permanently delete my account&amp;rdquo;&lt;/p&gt;</description></item><item><title>Operational Security: Staying safe online</title><link>https://ulveon-thoughts-f210db.gitlab.io/p/2024-10-05-operational-security-staying-safe-online/</link><pubDate>Sat, 05 Oct 2024 23:39:19 +0100</pubDate><guid>https://ulveon-thoughts-f210db.gitlab.io/p/2024-10-05-operational-security-staying-safe-online/</guid><description>&lt;p&gt;The Internet is a dangerous place. Especially if you are LGBTQ, furry, have unconventional sexual tendencies (CNC, BDSM, FinDom, etc) or are a member of a religious or racial minority group.&lt;/p&gt;
&lt;p&gt;The Internet was not designed with safety or security in mind. In its ARPANET beginnings, many systems didn&amp;rsquo;t even have passwords because only privileged users were supposed to access them. As such, privacy and security features are often little more than layers added on top of aging protocols, like HTTPS is just SSL applied on top of HTTP.&lt;/p&gt;</description></item></channel></rss>