<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 highly regulated sectors like healthcare. This tactic, detailed across multiple reports, creates a dual extortion threat: pay the ransom or face regulatory penalties 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 RCE 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 various 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 gain admin-level access to AWS GitHub repositories, including the JavaScript SDK, posing a supply chain risk compared to SolarWinds. AWS patched the issue within 48 hours and found no evidence of in-the-wild exploitation, according to Wiz and AWS statements. <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 features</strong>: GitHub Copilot now includes ‘memories’ in Visual Studio, allowing it to learn and document coding preferences and project guidelines, with a new cross-agent memory system enabling AI agents to share learned patterns and knowledge across development workflows. <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 surpasses Apple as top client</strong>: TSMC reported a 25.5% year-over-year revenue growth in Q4 2025, forecasting 30% growth for FY 2026, driven by persistent AI infrastructure demand, with Nvidia reportedly becoming its largest client, surpassing Apple. <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>: The Wikimedia Foundation has launched commercial partnerships with Amazon, Meta, Microsoft, Mistral AI, and Perplexity via Wikimedia Enterprise, monetizing 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>: Raspberry Pi launched the AI HAT+ 2 add-on board, equipped with a Hailo-10H neural network accelerator, enabling local execution of smaller LLMs and computer vision models on Raspberry Pi 5, eliminating cloud dependency. <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 rating downgrades. <a href="https://app.daily.dev/posts/9fgRHJtn9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quantum computing advances and post-quantum cryptography adoption</strong>: Hitachi Solutions Create launched DoMobile Ver.5 in Japan, integrating NIST-standardized post-quantum cryptography, while Quandela highlights hybrid quantum-classical systems for AI and progress in error correction as key 2026 trends. <a href="https://app.daily.dev/feed-by-ids?id=2U66SslD2&amp;id=DVTU5NL6H&amp;id=FzKtOLWjs&amp;id=MoM3tO9qY&amp;id=YjJQ2NnXc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Eurail BV customer database breached</strong>: Eurail BV disclosed a cybersecurity incident affecting its customer database, resulting in unauthorized access to personal information including names, contact details, and passport data, with DiscoverEU participants potentially having additional sensitive data exposed. <a href="https://app.daily.dev/posts/SkuYRBlp7" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-powered DevOps: the next big thing, or just more buzzwords?</strong>: Apparently, we now need ‘AI-powered DevOps’ courses, because traditional DevOps wasn’t complicated enough. Telusko is offering a four-month live course promising to integrate AI for productivity, which sounds suspiciously like adding ‘AI’ to everything to justify a higher price tag. One developer even claims to build production-ready apps in 2-3 hours with AI, shifting their role from coding to ‘reviewing and validating’ AI-generated code. So, if you’re still writing code, you’re clearly doing it wrong. <a href="https://app.daily.dev/feed-by-ids?id=SptJXtiZK&amp;id=QNz7Z2v2R" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Agentic AI: the new frontier of security risks, not solutions</strong>: Harmonic Security found 2.6% of 22.4 million GenAI prompts contained sensitive corporate data, with 71% of those exposures via ChatGPT. Meanwhile, the ‘real risk’ is apparently ‘workflow security,’ not the models themselves, as attackers are using Chrome extensions to steal ChatGPT data from 900,000 users. So, while companies are busy building ‘agentic memory systems’ for Copilot, the actual problem is users leaking data and malicious extensions, not some grand AI conspiracy. Classic misdirection. <a href="https://app.daily.dev/feed-by-ids?id=F9SzKFotn&amp;id=mZZUloj80&amp;id=940A3pjU5&amp;id=NCcbuPG8X&amp;id=NVsuaRFG5&amp;id=UAGIEGe4I" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft, Sony, Nintendo: still committed to player safety, five years later</strong>: Microsoft, Sony, and Nintendo have released a joint update on their ‘collaborative player safety initiative,’ first announced in 2020. Five years later, they’re still talking about ‘prevention, partnership, and responsibility’ without a single specific policy change or feature addition. It’s almost as if they’re just re-announcing their commitment to safety every few years to make it look like they’re doing something, while offering zero concrete details. Impressive dedication to the art of saying nothing new. <a href="https://app.daily.dev/posts/CTtMMpBVE" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
