<h2>Must know</h2>
<ul>
<li><strong>North Korean hackers use AI deepfakes for crypto attacks</strong>: North Korean threat actor UNC1069 is now using AI-generated deepfake videos, compromised Telegram accounts, and fake Zoom meetings to deploy multiple malware families targeting the cryptocurrency sector. This sophisticated social engineering campaign, detailed by Google Cloud, deploys new tools like SILENCELIFT and DEEPBREATH to harvest credentials and bypass macOS protections, marking a significant evolution in their tactics from traditional spear-phishing to AI-enabled deception. <a href="https://app.daily.dev/posts/wgv654rDT" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical Ivanti EPMM vulnerabilities exploited for dormant backdoors</strong>: Attackers are exploiting zero-day vulnerabilities (CVE-2026-1281, CVE-2026-1340) in Ivanti Endpoint Manager Mobile (EPMM) to plant dormant, in-memory backdoors. These ‘sleeper shells’ don’t touch disk and wait for specific triggers, making detection extremely difficult. The Dutch Data Protection Authority and Council for the Judiciary were already breached this way, exposing staff data. Organizations are advised to patch immediately and consider vulnerable instances compromised, as patching alone might not remove the in-memory implants, according to Hacker News. <a href="https://app.daily.dev/feed-by-ids?id=oF5bVmdnY&amp;id=Xu1XKqeJe&amp;id=bC7aQZbD9&amp;id=CZgK5a0B2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Discord to enforce global age verification</strong>: Discord is implementing mandatory global age verification by March 2026, treating all users as ‘teen by default’ until verification is completed via facial estimation or ID submission. This move, following previous UK/Australia rollouts that saw data breaches, is linked to lawsuits alleging minor exploitation and Discord’s upcoming IPO. Unverified accounts will face significant restrictions on sensitive content and direct messages from strangers. <a href="https://app.daily.dev/posts/J7jpcebnc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>EU accuses Meta of WhatsApp AI antitrust violation</strong>: The European Commission has formally accused Meta of violating EU competition rules by blocking third-party AI assistants like ChatGPT from WhatsApp, making Meta AI the sole option. <a href="https://app.daily.dev/posts/w9RLQf8YH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents develop C compiler for Linux kernel</strong>: Anthropic researchers used 16 Claude Opus 4.6 AI agents to autonomously build a Rust-based C compiler that successfully compiles the Linux kernel 6.9, though the generated code is inefficient and not production-ready. <a href="https://app.daily.dev/feed-by-ids?id=qhPl6UlEO&amp;id=9jEdtbewr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Salesforce shifts Heroku to sustaining engineering</strong>: Salesforce has moved its PaaS platform Heroku into a ‘sustaining engineering’ phase, halting new enterprise contract sales, which analysts interpret as a precursor to phasing out the service. <a href="https://app.daily.dev/feed-by-ids?id=13PGZg8yI&amp;id=smfFrov84" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel permanently adopts Rust</strong>: Linux 7.0 officially concludes the ‘Rust experiment,’ making Rust a permanent and supported language within the kernel project, signaling a long-term commitment. <a href="https://app.daily.dev/posts/k0qXAqGCH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google validates hard-braking events for crash prediction</strong>: Google Research validated that hard-braking events from Android Auto are reliable predictors of road crash risk, providing 18 times more data than traditional reports and enabling proactive identification of dangerous road segments. <a href="https://app.daily.dev/posts/HeAsnQEG7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI’s impact on open source software engagement</strong>: AI assistants are causing a decline in open source engagement, with developers using AI as an intermediary (‘vibe coding’) instead of directly interacting with projects, threatening OSS economic models and increasing maintainer burnout. <a href="https://app.daily.dev/feed-by-ids?id=0CiZDAswQ&amp;id=0hMgfxxrs&amp;id=7G6kpg12b&amp;id=8yRieynOP&amp;id=KCC82Mw6D&amp;id=d4YFQW1JT&amp;id=eD5zl44R8&amp;id=gVU0l8bpQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New DNS-PERSIST-01 validation method for certificates</strong>: A new DNS-PERSIST-01 domain control validation method is coming in 2026, allowing Certificate Authorities to issue certificates based on a single, static DNS record, significantly reducing complexity for certificate automation but raising security concerns. <a href="https://app.daily.dev/posts/7hFHkMPwg" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s Claude Co-work plugins ‘wiped out $400 billion in tech stocks’</strong>: Anthropic released Claude Co-work with industry-specific plugins, and apparently, this single event ‘wiped out $400 billion in tech stocks.’ Because a new set of plugins for an AI model is definitely the sole, definitive cause of such a massive market shift, and not, you know, the usual market volatility. The article’s title is a masterclass in definitive, uncorroborated market analysis. <a href="https://app.daily.dev/posts/QnkkrfiEd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will make traditional SaaS UIs irrelevant, says Databricks CEO</strong>: Databricks CEO Ali Ghodsi claims AI will render traditional SaaS UIs irrelevant through natural language interfaces, not by replacing systems of record. So, all those years building intuitive UIs? Wasted. Just talk to your computer, it’ll figure it out. Because natural language is always precise, unambiguous, and never leads to frustrating misunderstandings, especially when dealing with complex enterprise software. <a href="https://app.daily.dev/feed-by-ids?id=5kN0YzGRy&amp;id=BoDPkFq8Y&amp;id=cX7pVnhKE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI therapy chatbots are rising to address mental health</strong>: AI therapy chatbots like Wysa and Woebot are supposedly rising to address mental health needs. Meanwhile, an Oxford study found AI chatbots don’t improve medical advice compared to search engines and give inconsistent, even contradictory, recommendations. So, if you’re looking for inconsistent, potentially harmful advice, but delivered by an AI, then by all means, sign up. What could go wrong with automating mental health support? <a href="https://app.daily.dev/feed-by-ids?id=KxYxbzZtb&amp;id=eu7ShA5er&amp;id=jmvy6d70v&amp;id=nTkvD3DKT&amp;id=sOrxdsIQw" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
