<h2>Must know</h2>
<ul>
<li><strong>AWS EC2 now supports nested virtualization</strong>: AWS has finally rolled out nested virtualization support for its virtual EC2 instances, specifically on C8i, M8i, and R8i instances powered by Intel Xeon 6 processors. This means you can now run KVM or Hyper-V inside your EC2 VMs, a feature previously limited to bare metal. This brings AWS in line with Azure and Google Cloud, opening up new possibilities for testing multi-VM environments, running mobile emulators, or even Windows Subsystem for Linux on virtual workstations. <a href="https://app.daily.dev/posts/hFJsILiy4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hundreds of Chrome extensions caught stealing browsing data</strong>: A security researcher uncovered 287 Chrome extensions, with a staggering 37 million combined installs, secretly siphoning off users’ entire browsing histories. These extensions, masquerading as VPNs, productivity tools, and shopping add-ons, used encryption and obfuscation to hide their data exfiltration to external servers, likely for data brokers. This highlights a significant privacy and security risk within the browser extension ecosystem. <a href="https://app.daily.dev/posts/G9AfdVV0Z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>FTC launches antitrust probe into Microsoft’s cloud and AI practices</strong>: The FTC has launched an antitrust investigation into Microsoft, probing whether the tech giant is making it difficult or costly for customers to run its software on third-party cloud platforms. This inquiry also extends to Microsoft’s AI operations, licensing practices, and datacenter strategies, mirroring similar investigations in the UK and EU. This could have significant implications for cloud competition and how AI services are delivered. <a href="https://app.daily.dev/feed-by-ids?id=FEZjUlQd7&amp;id=FEZjUlQd7&amp;id=Zgule5sYq&amp;id=Zgule5sYq&amp;id=eOvUaW5L9&amp;id=eOvUaW5L9&amp;id=kwQJJqWna&amp;id=kwQJJqWna" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Istio updates bring security fixes and new features</strong>: Istio released versions 1.27.7 and 1.29.0, addressing critical security vulnerabilities and introducing new features like DNS proxying for ambient mesh and beta support for Gateway API Inference Extension. <a href="https://app.daily.dev/feed-by-ids?id=Hl8MqfAdH&amp;id=Vx9QllueY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Intel Arc B390 GPU shows strong Linux compute performance</strong>: Intel’s Arc B390 graphics, based on the Xe3 architecture in Panther Lake processors, shows strong OpenCL compute performance on Linux, outperforming previous generations and competing well against AMD’s Ryzen AI 9 HX 370. <a href="https://app.daily.dev/posts/0Bzv8LvNY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GlassFish 8.0 released with Jakarta EE 11 compatibility</strong>: GlassFish 8.0 is out, fully compatible with Jakarta EE 11, requiring JDK 21+, and featuring virtual thread support, Jakarta Data 1.0 integration, and enhanced security. <a href="https://app.daily.dev/posts/c8IbRHUTq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>CSS continues to evolve, reducing JavaScript reliance</strong>: Modern CSS has evolved to handle complex UI tasks like custom selects and scroll-driven animations, reducing the need for JavaScript and transforming CSS into a more powerful language for presentation and interaction. <a href="https://app.daily.dev/feed-by-ids?id=1TO1NBQ6P&amp;id=E3ttiNVjr&amp;id=G7unzK9jJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Alibaba’s Qwen 3.5 Plus model lands on Vercel AI Gateway</strong>: Alibaba’s Qwen 3.5 Plus model is now available on Vercel’s AI Gateway, featuring a 1M context window and excelling in agentic workflows, code generation, and multimodal tasks. <a href="https://app.daily.dev/feed-by-ids?id=4vqf4cHpB&amp;id=7m6C5PrGn&amp;id=8Nmwedcuv&amp;id=9Vcfez1Es&amp;id=UNO27M8rA&amp;id=XlMjalxp1&amp;id=dNY05UeZN&amp;id=fFNzT4nn1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS launches new AMD EPYC-powered M8azn instances</strong>: AWS has introduced new M8azn instances powered by fifth-gen AMD EPYC processors, offering double the compute performance and 4.3x memory bandwidth compared to M5zn instances. <a href="https://app.daily.dev/feed-by-ids?id=4vqf4cHpB&amp;id=7m6C5PrGn&amp;id=8Nmwedcuv&amp;id=9Vcfez1Es&amp;id=UNO27M8rA&amp;id=XlMjalxp1&amp;id=dNY05UeZN&amp;id=fFNzT4nn1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI chip design startup raises $335M</strong>: Ricursive Intelligence, founded by ex-Google Alpha Chip creators, raised $335M at a $4B valuation for AI tools that automate chip design, aiming to reduce design time from over a year to hours. <a href="https://app.daily.dev/feed-by-ids?id=FEZjUlQd7&amp;id=FEZjUlQd7&amp;id=Zgule5sYq&amp;id=Zgule5sYq&amp;id=eOvUaW5L9&amp;id=eOvUaW5L9&amp;id=kwQJJqWna&amp;id=kwQJJqWna" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents are ‘reputation farming’ in open source</strong>: Apparently, AI agents are now ‘reputation farming’ in open-source projects by submitting hundreds of pull requests, according to ‘security experts.’ This is supposedly a new threat for future supply chain attacks, unlike the XZ-utils backdoor which took years to build trust. So, instead of actual code, we’re now worried about AI bots getting GitHub stars. The future is truly here, and it’s… performative. <a href="https://app.daily.dev/posts/D9JdnEIyW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>KPMG partner caught using AI to cheat on AI exam</strong>: A KPMG partner got fined for using AI to cheat on an internal AI training exam. Because, you know, the best way to learn about AI is to have AI do your learning for you. This, along with over two dozen other KPMG staff caught doing the same, really instills confidence in their ‘AI expertise.’ Maybe they should just ask the AI to write their ethics policies too. <a href="https://app.daily.dev/feed-by-ids?id=79L5QoUWr&amp;id=IHwlDHiN7&amp;id=RAmInAYNd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is too dumb for my job, but smart enough for yours</strong>: The ‘GeLLMan Amnesia’ effect suggests people believe AI will automate everyone else’s job, but not their own, because they see the flaws up close. Meanwhile, ‘security experts’ are warning that LLM-generated malware is enabling ‘low-skilled operators’ to build exploitation frameworks. So, AI is simultaneously too dumb to take <em>my</em> job, but smart enough to empower script kiddies to become master hackers. Got it. <a href="https://app.daily.dev/feed-by-ids?id=6sTumutah&amp;id=6sTumutah&amp;id=FdIOjdD8s&amp;id=FdIOjdD8s&amp;id=hzvvYMQze" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
