<h2>Must know</h2>
<ul>
<li><strong>Microsoft patch tuesday addresses zero-days and AI vulnerabilities</strong>: Microsoft’s February 2026 Patch Tuesday fixed 59 vulnerabilities, including six actively exploited zero-days. Notably, three of these zero-days were publicly disclosed with potential proof-of-concept exploits. The update also patched critical remote code execution flaws in GitHub Copilot and multiple IDEs (VS Code, Visual Studio, JetBrains) stemming from AI prompt injection vulnerabilities, posing significant risks to developers due to access to sensitive credentials. Organizations should upgrade to the latest fixed versions following standard patching procedures. <a href="https://app.daily.dev/feed-by-ids?id=zk4Rm0PBv&amp;id=BOPPPXswR&amp;id=i2YjgAdJA&amp;id=oUCth2teJ&amp;id=uxvhPzT2r&amp;id=zFYu8Wov8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents introduce security risks and governance challenges</strong>: The rapid deployment of autonomous AI agents, expected in 40% of enterprise applications by 2026, brings significant security vulnerabilities and policy concerns. GitGuardian secured $50M to address the non-human identity crisis and AI agent security gaps, while Cisco introduced AI BOM for inventorying AI assets. Experts warn that competitive pressures are leading to rapid deployment without adequate safety frameworks, highlighting issues like prompt injection and the need for robust governance. <a href="https://app.daily.dev/feed-by-ids?id=0rQwEgnbB&amp;id=QRBfTNO1G&amp;id=ZVqULJLz3&amp;id=gUwIhEnZO&amp;id=l8YCDT7qf&amp;id=mmQEWnSC4&amp;id=UhlBeqTGk&amp;id=1X0JlaofS&amp;id=48BFhb5Uw&amp;id=73Kh0XGFs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source GLM-5 language model released</strong>: <a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a> released GLM-5, a new open-source large language model under an MIT License, claiming it matches proprietary models like Claude Opus 4.6 and GPT 5.3 Codex for long-horizon coding tasks. With 744B parameters and trained on 28.5T tokens, it integrates DeepSeek Sparse Attention and a novel asynchronous RL infrastructure. Modal offers free access via an OpenAI-compatible API endpoint, and the model integrates with popular AI frontends and coding agents. <a href="https://app.daily.dev/posts/bOspGaOnW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>ESLint 10.0.0 removes .eslintrc support</strong>: ESLint 10.0.0 will make flat config the sole configuration option, removing support for legacy .eslintrc files and improving config resolution for monorepos. <a href="https://app.daily.dev/posts/eQ0JUyyzE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vercel flags now in public beta</strong>: Vercel Flags offers native feature flag management directly in the Vercel Dashboard, including targeting rules, user segments, and SDK support for Next.js and SvelteKit. <a href="https://app.daily.dev/posts/IfJOyXyyM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>React Native 0.84 defaults to hermes v1</strong>: React Native 0.84 makes Hermes V1 the default JavaScript engine, bringing automatic performance improvements, reduced memory usage, and requiring Node.js 22.11 minimum. <a href="https://app.daily.dev/posts/oBV7ttAiR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VS Code 1.109 introduces AI-powered features</strong>: VS Code 1.109 adds AI-powered features like agent clarifying questions, MCP apps support for interactive UIs, and expanded AI model options through GitHub Copilot. <a href="https://app.daily.dev/posts/QJP16aVAU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WebMCP standard for AI agent interaction in chrome</strong>: WebMCP, a proposed web standard, enables websites to expose structured tools directly to in-browser AI agents, replacing brittle DOM scraping with explicit function calls. <a href="https://app.daily.dev/feed-by-ids?id=Ib422mkez&amp;id=NpNZ22j1b&amp;id=R0k8YNGTx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks launches lakeflow for data engineering</strong>: Databricks announced Lakeflow, a unified data engineering solution for ingestion, transformation, orchestration, and governance, featuring point-and-click data ingestion and simplified ETL. <a href="https://app.daily.dev/feed-by-ids?id=Bf1dK14FP&amp;id=DckO36Vaq&amp;id=cW6KgKgEK&amp;id=omSTWjEfj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Erlang/OTP 29.0-rc1 released with new features</strong>: Erlang/OTP 29.0-rc1 introduces native records, multi-valued comprehensions, a new <code>is_integer/3</code> guard BIF, and compiler improvements for binary operations and map comprehensions. <a href="https://app.daily.dev/posts/fTCktIrzF" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI will make you obsolete, unless you buy our tools</strong>: Apparently, AI models have reached an “inflection point” where they can complete complex tasks autonomously, meaning 50% of entry-level white-collar jobs are toast in 1-5 years. The solution? Immediately adopt paid AI tools and experiment daily, because the window for competitive advantage is closing. Or, you know, maybe it’s just another vendor trying to scare you into buying their stuff, like the article from The Register suggests about AI scaling challenges. <a href="https://app.daily.dev/feed-by-ids?id=jQIyW7C3o&amp;id=lGyFb7XTB&amp;id=waJpUtc5V&amp;id=mgWYZMYHx&amp;id=yCp2miBaq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Recruitment is hunger games, but here are some tips</strong>: Suzi Edwards-Alexander declares technical recruitment “fundamentally broken” and a “race to the bottom” driven by AI, comparing it to the Hunger Games. Her solutions include values-based interviews and focusing on “culture-add.” So, the system is broken, but don’t worry, just practice interviewing more and leverage your network. Because, clearly, the problem isn’t the system, it’s you not playing the game hard enough. <a href="https://app.daily.dev/posts/LD0ov49gc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will fix all your testing problems, just trust us</strong>: Datadog launched Code Coverage with AI-generated tests, and JiTTests promise to eliminate test maintenance costs by auto-healing tests. GitHub Copilot Testing for .NET is also generally available, offering AI-powered unit test generation. The article “The Death of Traditional Testing” even claims agentic development has broken traditional paradigms. So, apparently, AI is here to magically solve all our testing woes, because who needs human testers when you have algorithms that can infer intent and simulate faults? <a href="https://app.daily.dev/feed-by-ids?id=gZJqHPom7&amp;id=j254S16VC&amp;id=sJ5P2kSW8" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
