<h2>Must know</h2>
<ul>
<li><strong>Ransomware gangs weaponize compliance violations</strong>: Ransomware groups like Anubis and RansomHub are now threatening to report victims’ compliance violations to regulators, especially in healthcare. This tactic, highlighted by multiple security reports, creates a double bind: pay the ransom or face regulatory fines and reputational damage. Attackers are reportedly using AI to analyze stolen data for compliance gaps and even draft legal complaints, making this a rapidly evolving and potent threat. <a href="https://app.daily.dev/posts/OsQyB9bxn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cisco patches critical zero-day exploited by China-linked APT</strong>: Cisco has released emergency patches for a critical RCE vulnerability (CVE-2025-20393) in AsyncOS for Secure Email Gateway and Secure Email and Web Manager. This flaw, actively exploited by the China-linked APT group UAT-9686 since late 2025, allows arbitrary command execution as root if the Spam Quarantine feature is internet-exposed. Cisco’s fix also aims to remove persistence mechanisms installed by attackers, as reported by security outlets. <a href="https://app.daily.dev/posts/Faoj4Wg3E" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS CodeBuild webhook flaw exposed global cloud supply chain</strong>: Wiz security researchers uncovered ‘CodeBreach,’ a critical vulnerability in AWS CodeBuild stemming from improperly anchored regex patterns in webhook filters. This flaw could have allowed attackers to bypass security checks and gain admin-level access to AWS SDK repositories, potentially leading to global supply chain attacks. AWS patched the vulnerability within 48 hours of notification and found no evidence of in-the-wild exploitation, according to Wiz and other tech outlets. <a href="https://app.daily.dev/feed-by-ids?id=q7gLeN4EH&amp;id=MZnlHztN0&amp;id=PTMQTWkW2" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>GitHub Copilot gets agentic memory in Visual Studio</strong>: GitHub Copilot now features ‘memories’ that learn and document coding preferences and project guidelines, aiming to reduce code review friction and streamline onboarding, with an A/B test showing a 7% increase in pull request merge rates. <a href="https://app.daily.dev/posts/nCmOnsQS3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>TSMC revenue surges on AI demand, Nvidia becomes top client</strong>: TSMC reported 25.5% YOY revenue growth in Q4 2025, forecasting 30% growth for FY 2026, driven by persistent AI infrastructure demand, with Nvidia reportedly surpassing Apple as its largest client. <a href="https://app.daily.dev/posts/aTc9ls9yL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wikimedia Foundation partners with major AI companies</strong>: Wikimedia Foundation is now commercially partnering with Amazon, Meta, Microsoft, Mistral AI, and Perplexity via Wikimedia Enterprise to monetize AI companies’ use of Wikipedia data for training large language models. <a href="https://app.daily.dev/posts/QmEAqVcYI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Raspberry Pi releases AI HAT+ 2 for on-device generative AI</strong>: The new Raspberry Pi AI HAT+ 2, equipped with a Hailo-10H neural network accelerator and 8GB RAM, enables local execution of smaller LLMs and computer vision models, eliminating cloud dependency for edge AI applications. <a href="https://app.daily.dev/posts/nbjHwDlw3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oracle faces class action over undisclosed debt for OpenAI datacenters</strong>: Oracle is being sued by bondholders who allege the company failed to disclose an additional $38 billion in debt intended for datacenter investments supporting a $300 billion deal with OpenAI, leading to bond downgrades. <a href="https://app.daily.dev/posts/9fgRHJtn9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Eurail BV discloses customer database breach</strong>: Eurail BV announced a cybersecurity incident exposing personal information, including names, contact details, and passport data, with DiscoverEU participants potentially having additional sensitive data like IBANs and health information compromised. <a href="https://app.daily.dev/posts/SkuYRBlp7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Palo Alto Networks patches firewall DoS flaw</strong>: Palo Alto Networks issued emergency patches for CVE-2026-0227, a denial-of-service vulnerability in PAN-OS firewalls with GlobalProtect enabled, allowing unauthenticated remote attackers to force devices into maintenance mode. <a href="https://app.daily.dev/posts/MtreEtTLc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI is changing everything, except when it’s not</strong>: Apparently, AI is so transformative that it’s making junior developer jobs disappear because it can handle ‘boilerplate code,’ yet simultaneously, the real challenge has shifted from writing code to ‘understanding problems’ and ‘architecting systems.’ So, AI is both replacing coders and making coding less important, while still requiring ‘experienced oversight’ and ‘human judgment.’ It’s almost like the goalposts keep moving, but the hype train never stops. <a href="https://app.daily.dev/feed-by-ids?id=BlB3n2pFk&amp;id=LCRb60bYH&amp;id=pnpBRwBZD" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI security: because who needs actual security when you have ‘AI’?</strong>: Harmonic Security found 2.6% of 22.4 million prompts contained sensitive corporate data, mostly via ChatGPT, and 4% of that data ended up in China. Meanwhile, Trend Micro is touting an AI-powered platform, ÆSIR, that finds zero-day vulnerabilities ‘at the speed of AI.’ So, AI is both the problem (leaking sensitive data) and the solution (finding vulnerabilities), but let’s not talk about the fact that most attacks manipulate AI tool workflows, not the models themselves. It’s a security ‘risk sandwich,’ and we’re all apparently the filling. <a href="https://app.daily.dev/feed-by-ids?id=F9SzKFotn&amp;id=mZZUloj80&amp;id=Rj9ax7flK&amp;id=NCcbuPG8X&amp;id=NVsuaRFG5&amp;id=UAGIEGe4I" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oracle’s MySQL: open source in name only, apparently</strong>: According to some ‘community advocate sources,’ Oracle is maintaining MySQL’s ‘legal open source status’ while blocking community development and running all code changes privately. They’re also accused of releasing versions ‘plagued by severe technical issues’ and having ‘declining development activity.’ But don’t worry, MariaDB is the ‘true open source’ alternative. It’s almost like Oracle bought MySQL and then decided to slowly strangle it, all while pretending to be a good steward. Classic. <a href="https://app.daily.dev/feed-by-ids?id=WleLTp3IY&amp;id=kQKpHAixM" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
