<h2>Must know</h2>
<ul>
<li><strong>Microsoft’s patch tuesday addresses zero-days and AI vulnerabilities</strong>: Microsoft’s February 2026 Patch Tuesday is a doozy, fixing 59 vulnerabilities including six actively exploited zero-days. Critically, it also patches remote code execution flaws in GitHub Copilot and multiple IDEs (VS Code, Visual Studio, JetBrains) stemming from AI prompt injection. This highlights a new attack surface where AI tools, meant to help developers, can become vectors for significant security breaches, as reported by Krebs on Security. <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 new security risks and governance challenges</strong>: The rapid deployment of autonomous AI agents, expected in 40% of enterprise apps by 2026, brings significant security vulnerabilities and governance headaches. Cisco introduced AI BOM for inventorying AI assets, while GitGuardian secured $50M to tackle non-human identity (NHI) security and AI agent risks. Experts warn that competitive pressures are pushing rapid deployment without adequate safety frameworks, creating a tension between efficiency gains and potential exploitation, as highlighted by GitGuardian’s funding and Cisco’s new tool. <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>Databricks launches lakeflow for unified data engineering</strong>: Databricks unveiled Lakeflow on Azure Databricks, a comprehensive solution for data ingestion, transformation, orchestration, and governance. It features point-and-click ingestion, declarative ETL with automatic dependency mapping, and 99.9% reliable workflow orchestration. With serverless compute and AI-powered debugging, Lakeflow aims to streamline data engineering practices and provide unified governance through Unity Catalog. <a href="https://app.daily.dev/posts/DckO36Vaq" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Pandas 3.0 introduces breaking changes and copy-on-write</strong>: Pandas 3.0.0 is out with breaking changes, including a dedicated string dtype and formal Copy-on-Write semantics to finally kill view/copy ambiguity, requiring Python 3.11+. <a href="https://app.daily.dev/posts/Bf1dK14FP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source GLM-5 language model released by <a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a></strong>: <a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a> released GLM-5, a new open-source 744B parameter language model under an MIT License, claiming to match proprietary models like Claude Opus 4.6 and GPT 5.3 Codex for long-horizon coding tasks. <a href="https://app.daily.dev/posts/bOspGaOnW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Erlang/OTP 29.0-rc1 brings native records and quantum-resistant SSH</strong>: Erlang/OTP 29.0-rc1 introduces native records, multi-valued comprehensions, and defaults SSH to quantum-resistant key exchange, while dropping 32-bit Windows builds. <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. <a href="https://app.daily.dev/posts/5Mzv43k5F" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Amazon OpenSearch serverless introduces collection groups</strong>: Amazon OpenSearch Serverless now supports Collection Groups, allowing shared compute resources across multiple collections for cost reduction while maintaining security isolation. <a href="https://app.daily.dev/posts/5AicsK9Gn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS lake formation enhances cross-account sharing</strong>: AWS Lake Formation now supports enhanced cross-account sharing for hundreds of thousands of tables using a single AWS Resource Access Manager resource share with wildcard patterns. <a href="https://app.daily.dev/posts/5atbRAfWV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ClickHouse 26.1 brings deduplication and new indexing</strong>: ClickHouse 26.1 introduces end-to-end deduplication for asynchronous inserts, new indexing projection syntax, and faster DISTINCT operations, along with an open-source Kubernetes operator. <a href="https://app.daily.dev/posts/j8QVUmmhl" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-powered testing tools will kill traditional testing</strong>: Datadog, GitHub Copilot, and JiTTests are all pushing AI-powered testing tools, with one article even proclaiming ‘The Death of Traditional Testing.’ Apparently, LLMs will just magically generate and ‘auto-heal’ all your tests, eliminating maintenance costs. Because who needs actual human testers when you have an AI that can infer intent and simulate faults? What could possibly go wrong with letting an AI test the code written by another AI? <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>
<li><strong>AI will eliminate 50% of white-collar jobs in 1-5 years</strong>: An unnamed ‘industry leader’ is predicting 50% of entry-level white-collar jobs across law, finance, medicine, and software engineering will be gone in 1-5 years thanks to AI models like GPT-5.3 Codex and Claude Opus 4.6. This alarmist take suggests you should immediately adopt paid AI tools and experiment daily, or face obsolescence. Because, clearly, the only way to survive the AI apocalypse is to buy more AI tools. <a href="https://app.daily.dev/feed-by-ids?id=jQIyW7C3o&amp;id=lGyFb7XTB&amp;id=waJpUtc5V" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Recruitment is ‘fundamentally broken’ and ‘hunger games’</strong>: Suzi Edwards-Alexander claims technical recruitment is ‘fundamentally broken’ and a ‘race to the bottom’ driven by AI, calling it ‘What in the Hunger Games is Happening with Recruitment?’. While the process can be frustrating, framing it as a dystopian battle for survival might be a tad dramatic. Perhaps instead of blaming AI, we should just admit that hiring is hard and people are bad at it. <a href="https://app.daily.dev/posts/LD0ov49gc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
