<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 in multiple reports, creates a double bind: pay the ransom or face regulatory fines and reputational damage. They’re even using AI to analyze stolen data for compliance gaps and draft formal complaints, making this a rapidly evolving and serious 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 remote code execution (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 root-level command execution if the Spam Quarantine feature is internet-exposed. Cisco recommends immediate patching and notes the fix also removes attacker persistence, according to various security advisories. <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 major AWS GitHub repositories, including the JavaScript SDK, 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, as reported by 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>OpenAI invests in Altman’s brain-computer interface startup</strong>: OpenAI participated in a $250 million seed round for Merge Labs, a non-invasive brain-computer interface (BCI) startup co-founded by Sam Altman, raising potential conflict of interest concerns given Altman’s dual role. <a href="https://app.daily.dev/posts/rYkFwNLz3" target="_blank" rel="noopener nofollow">Read more</a></li>
<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, aiming to reduce code review friction and streamline onboarding by sharing learned patterns across AI agents. <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 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 surpassing Apple as TSMC’s 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>: 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 with 8GB onboard RAM. <a href="https://app.daily.dev/posts/nbjHwDlw3" 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 additional sensitive data exposed for DiscoverEU participants. <a href="https://app.daily.dev/posts/SkuYRBlp7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Agentic AI and non-human identity security challenges</strong>: The rapid adoption of AI-driven agentic systems for managing non-human identities (NHIs) in cloud environments is creating significant security blind spots and a ‘risk sandwich’ due to lagging security controls and the proliferation of ‘shadow agents.’ <a href="https://app.daily.dev/feed-by-ids?id=NCcbuPG8X&amp;id=NVsuaRFG5&amp;id=UAGIEGe4I" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Oracle’s debt drama: a $300 billion OpenAI deal, or just bad accounting?</strong>: Oracle is facing a class action lawsuit from bondholders who claim the company failed to disclose an additional $38 billion in debt, allegedly tied to a $300 billion deal with OpenAI. Apparently, Oracle’s bond offering documents were ‘misleading’ by omitting these massive datacenter investments. So, either Oracle forgot to mention a few tens of billions, or they’re just really bad at communicating their ‘multi-trillion-dollar arms race’ to build AI infrastructure. Transparency? Who needs it when you’re building the future! <a href="https://app.daily.dev/posts/9fgRHJtn9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Sensitive data exposure via GenAI: who’s surprised?</strong>: Harmonic Security found that 2.6% of 22.4 million prompts to generative AI tools contained company-sensitive data, with 71% of those exposures happening via ChatGPT. They’re warning that risks are amplified when this data trains public AI models, potentially leading to breaches that go undetected for months. It’s almost as if feeding proprietary information into a black box might not be the most secure strategy. Who could have seen that coming? <a href="https://app.daily.dev/feed-by-ids?id=F9SzKFotn&amp;id=mZZUloj80" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>MySQL’s open-source status: a legal fiction?</strong>: Oracle is maintaining MySQL’s ‘legal open source status’ while reportedly blocking community development and running all code changes privately, leading to declining development activity and severe technical issues like data corruption. Critics are calling its open-source status ‘nominal, not practical,’ suggesting MariaDB is the ‘true open source’ alternative. It seems Oracle’s idea of open source is more like a ‘look, don’t touch’ policy, which is great for community engagement, if you’re into one-way conversations. <a href="https://app.daily.dev/feed-by-ids?id=WleLTp3IY&amp;id=kQKpHAixM" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
