<h2>Must know</h2>
<ul>
<li><strong>AI agents are changing software engineering roles</strong>: AI coding agents are shifting the bottleneck in software development from writing code to defining requirements and understanding customer problems. This creates a new ‘Product Engineer’ role, blending technical skills with deep customer insight and cross-functional collaboration, as highlighted by Atlassian. Engineers now need strong communication and specification-writing skills to guide AI tools effectively, making soft skills essential and non-automatable, according to Hacker News. This means success depends on understanding the ‘why’ behind product decisions, not just the ‘what’.</li>
<li><strong>Java and Spring continue to evolve with AI integration</strong>: The Java and Spring ecosystems are actively integrating AI, with Spring AI and Quarkus with LangChain4j enabling production-grade agentic applications, as noted by Red Hat Developer. Oracle’s 2026 roadmap includes projects like Valhalla for performance and Loom for concurrency, enhancing Java’s capabilities for AI workloads. Meanwhile, new releases like Spring Framework 7.0.3 and WildFly 39 GA continue to refine the platform, ensuring Java remains a robust choice for enterprise AI, according to Java Weekly.</li>
<li><strong>AI coding tools introduce security vulnerabilities and architectural debt</strong>: While AI coding tools accelerate development, they introduce significant security vulnerabilities and architectural debt if not managed carefully. A Tenzai study found 69 vulnerabilities, many critical, in AI-generated code, particularly in business logic and API authorization. The New Stack and IT Revolution emphasize that AI-generated code often lacks architectural understanding, leading to increased bugs, churn, and review time, making human oversight and rigorous quality gates more critical than ever.</li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Small language models gain enterprise traction</strong>: Enterprises are increasingly favoring Small Language Models (SLMs) over Large Language Models (LLMs) due to lower costs, reduced latency, and better fit for specific, well-defined tasks like invoice parsing and contract classification, with models like Phi-3 and Mistral 7B matching LLM performance on constrained tasks.</li>
<li><strong>GitHub Copilot gets repository-specific memory</strong>: GitHub Copilot introduced a ‘memories’ feature in Visual Studio, allowing AI agents to learn and document coding preferences and project guidelines directly from user prompts and repository instructions, aiming to improve code consistency and speed up onboarding.</li>
<li><strong>Node.js and Ruby release security and bugfix updates</strong>: Node.js released versions 20.20.0 LTS and 25.3.0, addressing multiple CVEs and security vulnerabilities, while Ruby 4.0.1 was released with bugfixes, including a fix for spurious wakeups from Kernel#sleep.</li>
<li><strong>Microsoft open-sources XAML Studio for .NET UI prototyping</strong>: Microsoft has open-sourced XAML Studio under the .NET Foundation, providing a tool for rapid prototyping and fast feedback loops for XAML-based UI development, with features like a live XAML renderer and improved property inspection.</li>
<li><strong>Crystal 1.19.0 advances multithreading roadmap</strong>: Crystal 1.19.0 was released with execution contexts as a preview feature, a step towards enabling default multithreading in Crystal 1.20, alongside new synchronization primitives and ARM64 Linux binary support.</li>
<li><strong>Dynatrace enhances AI-driven log observability</strong>: Dynatrace upgraded its log observability platform with AI-powered contextual log analysis and unified data ingestion, adding features like Visual Resolution Path and customizable problem overviews to build trust in AI-driven root cause analysis.</li>
<li><strong>Arcjet releases Python SDK for application-layer security</strong>: Arcjet launched a Python SDK (beta) for application-layer security, embedding rate limiting, bot detection, and spam prevention directly into Python codebases for FastAPI and Flask, with code running locally via WebAssembly.</li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI coding assistants: the new treadmill for developers</strong>: Apparently, AI coding assistants are making developers ‘slower’ and generating ‘46% more code’ with ‘1.7x more bugs,’ according to The Serious CTO. So, while you’re feeling like a 10x developer, you’re actually just churning out more garbage for your colleagues to clean up. It’s like a gym membership where you pay to run in place, but with more security vulnerabilities and architectural debt.</li>
<li><strong>The ‘Product Engineer’: because ‘software engineer’ wasn’t vague enough</strong>: With AI handling the mundane task of actually writing code, engineers are now expected to become ‘Product Engineers,’ deeply understanding customer problems and validating solutions. So, instead of just coding, you’ll be doing user interviews, reviewing support tickets, and basically becoming a product manager who can also code. Because, you know, AI can’t do empathy… yet. This is according to Atlassian, which probably has a new tool to sell you for this ‘essential’ new role.</li>
<li><strong>AI governance: the new ‘battleground’ for IT leaders</strong>: Just when you thought you had a handle on shadow IT, now we have ‘shadow AI’ to worry about. Apparently, AI tools adopted outside formal oversight are creating ‘identity collapse’ and ‘permission creep,’ according to The New Stack. So, get ready for another round of ‘governance frameworks’ and ‘identity propagation’ for your AI agents, because apparently, we can’t trust our machines to behave without a mountain of bureaucracy.</li>
</ul>
