<h2>Must know</h2>
<ul>
<li><strong>Microsoft’s patch tuesday addresses zero-days and AI vulnerabilities</strong>: Microsoft’s February 2026 Patch Tuesday fixed 59 vulnerabilities, including six actively exploited zero-days. Critically, it also patched remote code execution flaws in GitHub Copilot and multiple IDEs (VS Code, Visual Studio, JetBrains) stemming from AI prompt injection vulnerabilities. These AI-related flaws pose significant risks to developers due to potential access to sensitive credentials and infrastructure keys, as reported by Krebs on Security. <a href="https://app.daily.dev/feed-by-ids?id=zk4Rm0PBv&amp;id=uxvhPzT2r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents introduce new 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, which just raised $50M, is focusing on securing these ‘Non-Human Identities’ and addressing gaps in AI agent security. Experts warn that competitive pressures are leading to rapid deployment without adequate safety frameworks, with prompt injection becoming a critical attack vector. <a href="https://app.daily.dev/feed-by-ids?id=gUwIhEnZO&amp;id=l8YCDT7qf&amp;id=UhlBeqTGk&amp;id=BOPPPXswR&amp;id=lhjmWCnek&amp;id=1X0JlaofS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source GLM-5 model claims parity with proprietary AI for coding</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 40B active, it integrates novel asynchronous RL infrastructure and is available for free access via Modal until April. This could significantly impact the landscape for AI-assisted development, offering a powerful open alternative. <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>Databricks launches lakeflow for unified data engineering</strong>: Databricks introduced Lakeflow on Azure Databricks, a comprehensive solution for data ingestion, transformation, orchestration, and governance, featuring AI-powered debugging and serverless compute. <a href="https://app.daily.dev/posts/DckO36Vaq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pandas 3.0.0 released with breaking changes and performance focus</strong>: Pandas 3.0.0 is out, requiring Python 3.11+, introducing a dedicated string dtype, formal Copy-on-Write semantics, and early support for declarative column transformations, aiming to address performance and clarity issues. <a href="https://app.daily.dev/posts/Bf1dK14FP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Erlang/OTP 29.0-rc1 brings new features and breaking changes</strong>: Erlang/OTP 29.0-rc1 introduces native records, multi-valued comprehensions, and quantum-resistant SSH key exchange, while removing 32-bit Windows builds and changing the default code path. <a href="https://app.daily.dev/posts/fTCktIrzF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>FastAPI cloud simplifies deployment with single command</strong>: FastAPI Cloud, backed by Sequoia, offers a new platform for deploying FastAPI applications with a single command, handling HTTPS, auto-scaling, and Kubernetes infrastructure, currently in private beta. <a href="https://app.daily.dev/posts/5Mzv43k5F" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>React Native 0.84 defaults to hermes v1 for performance</strong>: React Native 0.84 now uses Hermes V1 as 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>AWS expands EC2 instance availability with custom intel xeon 6 processors</strong>: AWS has expanded the regional availability of EC2 C8id, M8id, and R8id instances, powered by custom Intel Xeon 6 processors, offering up to 43% higher performance and 3.3x more memory bandwidth than previous generations. <a href="https://app.daily.dev/posts/eOlZw3FkV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vercel flags now in public beta for feature management</strong>: Vercel Flags is now in public beta, providing native feature flag management directly in the Vercel Dashboard with targeting rules, user segments, and environment controls, supporting Next.js and SvelteKit. <a href="https://app.daily.dev/posts/IfJOyXyyM" 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 my tools</strong>: An ‘industry leader’ is screaming that 50% of white-collar jobs will vanish in 1-5 years thanks to AI, urging everyone to immediately adopt paid AI tools and experiment daily. Because, clearly, the only way to survive the AI apocalypse is to buy their stuff and frantically keep up with the latest flavor-of-the-month model. The real ‘singularity’ is the one where everyone’s wallet gets lighter. <a href="https://app.daily.dev/posts/lGyFb7XTB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are the new microservices, just trust us</strong>: Microsoft Foundry is now an ‘end-to-end enterprise AI agent platform’ with a ‘unified successor’ to Semantic Kernel and AutoGen. They’re pushing multi-agent workflow orchestration and a ‘centralized Control Plane’ like it’s the second coming of microservices. Because what we really needed was another layer of abstraction and complexity, now with added AI-powered hallucinations and security risks. <a href="https://app.daily.dev/feed-by-ids?id=dz3njIXXE&amp;id=it7NMfnEb&amp;id=laJOXndNn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The death of traditional testing, brought to you by AI</strong>: Apparently, ‘agentic development has broken traditional testing paradigms,’ and now we have ‘JiTTests’ (Just-in-Time Tests) that are LLM-generated and ‘auto-heal’ themselves. GitHub Copilot Testing for .NET is also generally available, promising AI-powered unit test generation. So, just let the AI write your code, then let the AI test your code. What could possibly go wrong? Human testers, your job is safe… for now, until the AI decides it doesn’t need you to review its ‘true positives’. <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>
