<h2>Must know</h2>
<ul>
<li><strong>AI code generation: faster code, slower everything else</strong>: AI code generation can speed up coding by 55%, but it shifts bottlenecks to verification, increasing review times by 91% in high-AI-usage teams. This leads to production incidents, code churn, and a decline in code health, especially in legacy systems where AI lacks context. Companies are finding that investing in making knowledge AI-accessible can have a 4-month payback period, but without proper guardrails, AI just makes a mess faster, according to multiple reports. <a href="https://app.daily.dev/feed-by-ids?id=CUMsxmpoK&amp;id=H1aWwdchu&amp;id=KYToiOjQs&amp;id=TCb4dVj23&amp;id=k9rFokMvG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes ecosystem sees major shifts</strong>: The cloud native ecosystem is undergoing significant changes, with the Ingress NGINX project announcing its retirement and the Kubernetes Dashboard project being archived. On a brighter note, a new open-source Kubernetes dashboard called Radar has been introduced, built as a single Go binary with an embedded React frontend, offering real-time cluster visibility without cloud logins or agents. This aims to address limitations of existing tools like Lens and k9s. <a href="https://app.daily.dev/feed-by-ids?id=EdFbm39PB&amp;id=DiEunA7dc&amp;id=a3wJg58hN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI security and privacy concerns escalate</strong>: A staggering 175,108 Ollama AI hosts are exposed to the internet, running models without safety guardrails, creating potential for resource hijacking and remote attacks. SpaceX updated Starlink’s privacy policy to use customer data for AI training, coinciding with talks to merge with xAI. Additionally, a critical 1-click RCE vulnerability was found in OpenClaw, an AI assistant used by over 100,000 developers, highlighting the urgent need for secure AI agent architectures. <a href="https://app.daily.dev/feed-by-ids?id=2SdHHyeRJ&amp;id=YUSdAM3D6&amp;id=d76FYhLFT&amp;id=iEIaVGrhd&amp;id=yXHHUa4yo&amp;id=4nlzVZMEL&amp;id=ALGXJz29U&amp;id=fuL6WfKHt&amp;id=oAoGgXAAn&amp;id=zgJdriYuf" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>MongoDB formally verifies distributed transactions</strong>: MongoDB’s team formally verified its distributed transaction protocol, ensuring ACID guarantees and snapshot isolation across sharded clusters using compositional TLA+ specifications and a novel ‘permissiveness metric’ to analyze concurrency efficiency. <a href="https://app.daily.dev/posts/KMiaK5RJr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New PostgreSQL extension for distributed tracing</strong>: DataDog released <code>pg_tracing</code>, a PostgreSQL extension for server-side distributed tracing that generates spans for queries, functions, and transactions, supporting context propagation and OTLP export. <a href="https://app.daily.dev/posts/Xfr0u4GSj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wix introduces xEngineer model for AI-driven development</strong>: Wix launched its ‘xEngineer’ model, integrating AI to empower engineers with broader ownership across the development lifecycle, aiming to evolve responsibilities rather than reduce headcount. <a href="https://app.daily.dev/posts/1ESLU5xBs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WebAssembly celebrates 10 years, expands beyond browser</strong>: WebAssembly (WASM) marked its 10th anniversary, evolving from a browser performance solution to a foundational technology for edge platforms, serverless runtimes, and embedded systems, now showing promise for AI workloads. <a href="https://app.daily.dev/posts/XjOBMuNCm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel 6.19 extends release cycle</strong>: The Linux 6.19 release cycle was extended by a week due to holiday downtime, with the stable version now expected next week, and will include a new Uniwill laptop platform driver for improved hardware integration. <a href="https://app.daily.dev/feed-by-ids?id=WHsGYq8Jt&amp;id=mzBRRS79C" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New cloud VM services for AI-era needs</strong>: Three new cloud VM services (Sprites, exe.dev, shellbox) are gaining traction for providing fast sandboxed environments for LLM coding agents and zero-friction web service deployment, despite questionable price-to-performance ratios. <a href="https://app.daily.dev/posts/lPMG4kfe9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Feature-driven architecture for scalable applications</strong>: Organizing code around business features rather than technical layers, treating each feature as a self-contained unit, reduces cognitive overhead and enables parallel development for scalable applications. <a href="https://app.daily.dev/feed-by-ids?id=IZ8YVjtZx&amp;id=Y9sZHn6NM" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-generated video series: another flop for mainstream audiences</strong>: Director Darren Aronofsky and TIME magazine released an AI-generated video series, ‘On This Day… 1776,’ using Google DeepMind tech. Unsurprisingly, it received widespread criticism and negative reception from mainstream audiences, continuing a pattern of AI-generated visuals failing to resonate. Apparently, even with a big name attached, AI still can’t quite capture that human touch, or maybe it’s just that nobody asked for it. <a href="https://app.daily.dev/feed-by-ids?id=0hvZGEXEr&amp;id=vgyq27hp2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Native PHP promises free mobile development, powered by AI magic</strong>: Native PHP just dropped ‘major updates’ making mobile development ‘completely free and open source,’ complete with a slimmed-down core, native UI components, and an AI tool named Mimi that ‘rapidly generates complete Laravel mobile applications’ from natural language. It’s all very promotional, of course, and completely uncorroborated. Because nothing says ‘free and open source’ like a magic AI wand that builds your app for you, right? <a href="https://app.daily.dev/posts/bRf9BXT9O" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI delegation: MBAs building startups without coding, because management is the new tech</strong>: Apparently, MBA students, ‘lacking coding skills,’ successfully built startup prototypes in four days using AI tools. The takeaway? Management skills are now ‘critical for effective AI agent interaction,’ and ‘knowing what to ask for’ is the new scarce resource. So, while engineers are busy debugging AI-generated spaghetti code, the MBAs are out there ‘delegating’ their way to startup glory. Effort is abundant and cheap, but good luck finding someone to actually <em>do</em> the work. <a href="https://app.daily.dev/posts/WyavXiw9C" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
