<h2>Must know</h2>
<ul>
<li><strong>AI reaches inflection point, developers feel the heat</strong>: AI coding tools, like Claude Opus 4.5/4.6 and GPT-5.2/5.3, are now generating working, tested software from detailed specs. This has led to significant psychological distress among developers, termed ‘Deep Blue’ on the Oxide and Friends podcast, as many feel their skills are being devalued. While AI speeds up code generation, overall productivity isn’t necessarily improving due to new overheads like debugging AI-introduced bugs, as reported by SonarSource. <a href="https://app.daily.dev/feed-by-ids?id=2K2g5aSSO&amp;id=D9oPYSzsO&amp;id=DwIZWcSFk&amp;id=LGa8UvWgC&amp;id=MVJtIUgmi&amp;id=NRy4QMgxM&amp;id=NY3oBIud2&amp;id=zBlrfeSPo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New RCE and SSJI vulnerabilities found in Notepad and Lambda</strong>: Multiple critical vulnerabilities have been discovered, including a Remote Code Execution (RCE) in modern Windows Notepad versions (CVE-2026-20841) due to improper sanitization of <code>file://</code> URIs in Markdown links. Another, React2Shell (CVE-2025-55182), affects serverless Lambda deployments running Next.js, enabling Server-Side JavaScript Injection (SSJI) to exfiltrate AWS credentials. These highlight how new features and architectural differences can introduce unexpected attack surfaces. <a href="https://app.daily.dev/feed-by-ids?id=AAuuxbltu&amp;id=fSewiiKKS&amp;id=jk8Zg0MX3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Securing non-human identities in the cloud</strong>: Non-Human Identities (NHIs), such as API keys and tokens, are increasingly critical security assets in cloud environments. Advanced secrets management and NHI management are being pushed as solutions to secure machine identities, with claims of a 35% reduction in breach risk and improved compliance like SOC2. These practices aim to automate secrets rotation and enhance visibility, particularly benefiting finance, healthcare, and DevOps sectors. <a href="https://app.daily.dev/feed-by-ids?id=ISqrlUoED&amp;id=Ypw3CG2Vn" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Vulkan 1.3 now on Apple devices with KosmicKrisp</strong>: KosmicKrisp, a Vulkan-on-Metal driver developed by LunarG, has achieved feature parity with MoltenVK and been upstreamed to Mesa 26.0, enabling Vulkan 1.3 compliance on Apple devices. <a href="https://app.daily.dev/posts/2qlomA9u7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>India becomes OpenAI’s second-largest market</strong>: OpenAI CEO Sam Altman announced that India is now OpenAI’s second-largest market with 100 million weekly active ChatGPT users, driven by adaptations like a sub-$5 tier and free year-long subscriptions. <a href="https://app.daily.dev/posts/E7X2UrHD0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ruby toolkit for LLM integration released</strong>: A new Ruby toolkit, <code>llm.rb</code>, offers a zero-dependency solution for integrating multiple LLMs like OpenAI, Gemini, and Anthropic, providing a unified API for chat completions, streaming, and structured outputs. <a href="https://app.daily.dev/posts/eI4XzTrQw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source educational robot kit Reachi Mini</strong>: Hugging Face and Pollen Robotics released the Reachi Mini, an open-source educational robot kit powered by Raspberry Pi, offering local control via a Python SDK despite some setup ‘rough edges’ and privacy concerns with cloud connectivity. <a href="https://app.daily.dev/posts/f1KNRZS82" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Dash robot reverse-engineered for enhanced control</strong>: A developer successfully reverse-engineered the Dash educational robot, using Ghidra to uncover its BLE communication protocol and JSON message structure, enabling enhanced Python control libraries and discovering previously unknown autonomous behaviors. <a href="https://app.daily.dev/posts/GRUwDEkm4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>PERFECT framework for code reviews</strong>: The PERFECT framework for code reviews prioritizes Purpose, Edge cases, Reliability, Form, Evidence, Clarity, and Taste to reduce cognitive load and improve quality through clear conventions and automated checks. <a href="https://app.daily.dev/posts/XgjRdXZ44" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Startups prioritize automation over hiring speed</strong>: Startups are increasingly gaining competitive advantage by automating internal functions rather than just hiring faster, shifting strategy towards operational efficiency through technology. <a href="https://app.daily.dev/posts/iqSQYY6BU" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI floods cURL bug bounty with ‘slop’</strong>: Daniel Stenberg, the creator of cURL, reports that AI-generated security reports have flooded their bug bounty program, dropping accuracy from 1 in 6 to 1 in 20-30, forcing its shutdown. Apparently, AI is great at generating noise, but not so much at actual signal. Who knew that quantity doesn’t always equal quality? <a href="https://app.daily.dev/feed-by-ids?id=7e4x0vONT&amp;id=UMatWajxo&amp;id=rh5pHHgby" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI safety concerns at xAI lead to engineer exodus</strong>: Elon Musk’s xAI is reportedly losing engineers over concerns about AI safety, with sources claiming Musk wants Grok to be ‘more unhinged.’ Because nothing says ‘responsible AI development’ like actively pushing for less safety and more chaos, especially after your AI generates a million deepfake images. Clearly, the future of AI is less about intelligence and more about unchecked ego. <a href="https://app.daily.dev/posts/uhJEF1Wo4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hugo migration fails due to slow build times</strong>: A developer tried to migrate their blog from Jekyll to Hugo, only to find that Hugo’s external processing of Asciidoc files ballooned build times from 2 minutes to over 10 minutes for 800+ posts. So much for the promise of faster static site generation. Sometimes, the ‘modern’ solution just means more waiting, proving that new isn’t always improved. <a href="https://app.daily.dev/posts/SfIlGfEoG" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
