<h2>Must know</h2>
<ul>
<li><strong>AI agents are farming open-source reputation</strong>: An AI agent named Kai Gritun submitted over 100 pull requests across nearly 100 open-source repositories in a single day. Security experts are calling this ‘reputation farming,’ a new tactic to gain influence in critical infrastructure projects, potentially enabling future supply chain attacks. This method rapidly accumulates trust, unlike previous attacks that took years, and shifts the attack surface from code to governance processes, as reported by various security outlets. <a href="https://app.daily.dev/posts/D9JdnEIyW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS now supports nested virtualization on EC2</strong>: AWS has rolled out support for nested virtualization on virtual EC2 instances, a feature previously limited to bare metal. This allows users to run Hyper-V or KVM inside their EC2 instances, enabling use cases like mobile app emulators, automotive hardware simulation, and testing multi-VM environments. This brings AWS in line with Azure and Google Cloud, offering more flexibility for complex virtualized workloads on Intel Xeon 6-powered instances. <a href="https://app.daily.dev/posts/hFJsILiy4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Chrome zero-day actively exploited in the wild</strong>: Google released an emergency Chrome update to patch CVE-2026-2441, a high-severity use-after-free zero-day in CSS that is actively being exploited. This marks Chrome’s first zero-day of 2026, following eight in 2025. Security experts emphasize the critical need for rapid patching and endpoint monitoring, given the inevitability of vulnerabilities in Chrome’s massive codebase. <a href="https://app.daily.dev/feed-by-ids?id=AK3a3hE0z&amp;id=EWohDzjzG&amp;id=Rk6iVnRuo&amp;id=bvXFp0lun&amp;id=3jxdGuYhW&amp;id=8Y407Qiao&amp;id=9vJqBcvBu&amp;id=Cde5RaWOD&amp;id=JgDEYLcS1&amp;id=KCoKWf8AT" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI agents building a SQLite-like database in Rust</strong>: An experiment used multiple AI agents (Claude, Codex, Gemini) in parallel to build a SQLite-like database engine in Rust, generating ~19k lines of code with 282 passing tests, highlighting the potential for parallel AI agent development with proper coordination. <a href="https://app.daily.dev/feed-by-ids?id=pbRQZ4xYf&amp;id=Btt27SjiL&amp;id=Nvd4waKyi&amp;id=l9nJMN2FZ&amp;id=trqJdFpBG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Istio releases security patches and new features</strong>: Istio released versions 1.27.7 and 1.29.0, addressing high-severity security vulnerabilities like code smuggling in cgo binaries and TLS session resumption flaws, while also introducing new features like DNS proxying for ambient mesh and 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>WebAssembly running in WebGPU</strong>: An experimental project demonstrates running WebAssembly instructions on WebGPU by transpiling WASM to WGSL shaders, enabling C/C++ features for shader development and achieving reasonable performance. <a href="https://app.daily.dev/posts/FYyRebYAL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI model optimizes protein drug development</strong>: MIT researchers developed an AI model using LLMs to optimize codon sequences for protein production in industrial yeast, potentially cutting biologic drug development costs by 15-20% by improving engineering predictability. <a href="https://app.daily.dev/posts/9Vcfez1Es" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Intel Arc B390 GPU shows strong Linux performance</strong>: Intel’s Arc B390 graphics, based on Xe3 architecture within Panther Lake processors, demonstrates strong GPU compute performance on Linux using open-source drivers, outperforming previous Intel generations and competing with AMD’s Ryzen AI 9 HX 370 in OpenCL benchmarks. <a href="https://app.daily.dev/posts/0Bzv8LvNY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NetEase divests Anchor Point Studios</strong>: Anchor Point Studios, founded by NetEase, is transitioning to independent operation, seeking new investors for its debut action-adventure project, continuing NetEase’s trend of separating from or shuttering Western studios despite overall revenue growth. <a href="https://app.daily.dev/posts/tSheuurkN" 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>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI will make you a high-paid full stack developer, apparently</strong>: A ‘comprehensive 2026 roadmap’ claims that to be a high-paid full stack developer, you just need to integrate AI skills, use Python for LLMs, and shift focus from syntax to ‘product strategy’ because AI handles code generation. So, just wave your magic AI wand, and presto, you’re rich. Sounds legit. <a href="https://app.daily.dev/posts/JT1zQVK71" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The side project apocalypse is upon us</strong>: The ‘dream’ of a profitable side-project SaaS is dead, according to some, because AI-generated competitors will flood every niche faster than you can ship. Quality no longer matters, and the only winners are datacenter providers. So, if you’re not building for enterprise, just give up now, I guess. Or maybe, just maybe, build something actually good? <a href="https://app.daily.dev/posts/dIaDDgpAV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will automate everyone else’s job, not 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. The article reassures us that LLMs ‘never quite replace deep domain expertise.’ So, your job is safe, but your neighbor’s isn’t. Convenient, isn’t it? <a href="https://app.daily.dev/feed-by-ids?id=FdIOjdD8s&amp;id=6sTumutah" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
