<h2>Must know</h2>
<ul>
<li><strong>Claude Code adds parallel agents and batch migrations</strong>: Anthropic has updated its Claude Code CLI with two new skill agents, <code>/simplify</code> and <code>/batch</code>, to handle complex refactoring. The <code>/batch</code> command uses git worktrees to run dozens of isolated agents in parallel, each capable of managing its own files and submitting Pull Requests. According to the release notes, a new <code>/copy</code> command and HTTP hook handlers for external service integration have also been added to the toolset. <a href="https://app.daily.dev/posts/sq7rTuWl2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical ReDoS vulnerabilities found in minimatch</strong>: Three high-severity vulnerabilities (CVE-2026-27904, 27903, 26996) were patched in the <code>minimatch</code> npm package, which sees nearly 500 million weekly downloads. These flaws involve catastrophic backtracking in regex patterns that can stall the Node.js event loop for several seconds. Socket has released free certified patches that can be applied via their CLI without requiring a full dependency upgrade, as reported by security researchers. <a href="https://app.daily.dev/feed-by-ids?id=2WnWCE8iO&amp;id=89yzay3k6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Netflix solves container launch bottlenecks on modern CPUs</strong>: Netflix engineers identified a severe performance regression when migrating to <code>containerd</code> with user namespaces, causing 30-second timeouts on high-core-count Intel systems. Profiling revealed that kernel idmap mounts were competing for global VFS locks, a bottleneck amplified by NUMA memory latency. A fix contributed upstream to <code>containerd</code> reduces mount operations from O(n) to O(1), eliminating the lock contention entirely. <a href="https://app.daily.dev/posts/cKVVAN6ve" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Linux 7.1 patches introduce multi-scheduler support and socket extended attributes</strong>: The latest kernel updates include HR tick timer optimizations and a performance boost for Nouveau Vulkan Zcull. <a href="https://app.daily.dev/posts/T0tOfwEhy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>FreeBSD begins pushing Rust into its kernel</strong>: The project is officially integrating Rust components to improve memory safety within the core operating system. <a href="https://app.daily.dev/posts/T0tOfwEhy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Systemd v260 officially drops support for SysV init scripts</strong>: The long-deprecated legacy initialization scripts are finally being removed in the upcoming major release. <a href="https://app.daily.dev/posts/T0tOfwEhy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Metamorphic testing uncovers 21 bugs in zero-knowledge compilers</strong>: Researchers used the MTZK framework to find flaws in industrial ZK compilers that could lead to invalid proof acceptance. <a href="https://app.daily.dev/posts/EvvwDJaxE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gleam programming language gains traction for its Erlang and JS targets</strong>: The first Gleam conference highlighted the language’s simplicity and upcoming official guides for Postgres and auth integration. <a href="https://app.daily.dev/posts/saEXLDYMU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Argo CD 3.3 introduces PreDelete hooks for safer teardowns</strong>: The update enables predictable resource deletion sequences and adds shallow Git cloning to speed up monorepo fetches. <a href="https://app.daily.dev/posts/J931yCKTd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gemini 3 Deep Think targets complex architectural debugging</strong>: The new model variant is optimized for deep reasoning tasks and high-level software design decisions. <a href="https://app.daily.dev/posts/INaqwNVxY" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The AI-RAN hype train arrives at MWC</strong>: NVIDIA and partners are shouting about ‘AI-native 6G’ and ‘agentic AI workloads’ running on 5G base stations. While 36 Gbps throughput sounds impressive, most of this is just moving standard signal processing to GPUs so they can sell more GH200 servers. It’s a classic case of finding a problem (telecom infrastructure) for their very expensive hardware solution. <a href="https://app.daily.dev/feed-by-ids?id=N9cHfOw6g&amp;id=PJftzADbd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare rebuilt a framework in a week using AI</strong>: Cloudflare claims they used AI agents to clone a Next.js-compatible framework by just pointing them at the test suites. The Primeagen and others correctly point out that it lacks critical features like static pre-rendering, making it more of a ‘test-suite-passing demo’ than a production-ready tool. It’s a great marketing stunt for AI agents, but don’t expect to migrate your enterprise apps to it by Friday. <a href="https://app.daily.dev/posts/5XUEq41aC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenClaw is a ban speedrun for your AI accounts</strong>: OpenClaw is being marketed as an ‘open-source AI automation tool’ for your browser, but users are reporting immediate account bans from Google and Anthropic. Running automated scripts against proprietary LLM interfaces without an API is the fastest way to lose your access. It’s less of a productivity tool and more of a sophisticated way to get blacklisted by every major AI provider. <a href="https://app.daily.dev/posts/INaqwNVxY" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
