<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. Critically, it also addressed 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=BOPPPXswR&amp;id=uxvhPzT2r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source sustainability crisis and maintainer burnout</strong>: The open-source software ecosystem is facing a sustainability crisis, with companies profiting from the unpaid labor of maintainers who are experiencing burnout and a lack of support. High-profile incidents like XZ Utils highlight the critical dependencies maintained by overworked individuals. Lotte Pitcher at NDC London 2026 emphasized that consumers must reframe their support for OSS as risk reduction, not charity, by sponsoring projects and contributing time. <a href="https://app.daily.dev/posts/aIx715a23" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent security and non-human identity crisis</strong>: GitGuardian secured $50M in Series C funding to address the growing security challenges posed by AI agents and non-human identities (NHI). With 350,000 secret exposures detected in 2025, the company aims to innovate in AI agent security and enterprise-scale NHI governance. This highlights the critical need for robust security frameworks as AI agents gain more autonomous capabilities and access to sensitive systems, as reported by GitGuardian. <a href="https://app.daily.dev/feed-by-ids?id=UhlBeqTGk&amp;id=gUwIhEnZO&amp;id=BOPPPXswR&amp;id=i2YjgAdJA&amp;id=lhjmWCnek" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Erlang/OTP 29.0-rc1 released with native records and quantum-resistant SSH</strong>: Erlang/OTP 29.0-rc1 introduces native records, multi-valued comprehensions, and defaults to quantum-resistant <code>mlkem768x25519-sha256</code> key exchange for SSH, 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>Linux kernel 7.0 scheduler and performance improvements</strong>: Linux 7.0 merges significant scheduler improvements, including time slice extension, performance optimizations for high core count systems, and a 12% UDP receive performance boost on 100Gbit NICs by inlining <code>timecounter_cyc2time()</code>. <a href="https://app.daily.dev/feed-by-ids?id=RL4bMHXJh&amp;id=feLrRFW3p&amp;id=mdQGXm2CW&amp;id=s1iSS5CBn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks launches lakeflow for unified data engineering</strong>: Databricks introduced Lakeflow on Azure Databricks, a unified data engineering solution for ingestion, transformation, orchestration, and governance, featuring point-and-click ingestion, declarative pipelines, and AI-powered debugging. <a href="https://app.daily.dev/posts/DckO36Vaq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Modular monoliths balance complexity and operational ease</strong>: Modular monoliths offer microservices-level logical separation with monolith-level operational simplicity by treating modules as bounded contexts within a single deployable unit, enforcing architectural boundaries at build time. <a href="https://app.daily.dev/feed-by-ids?id=AaHqpuNEs&amp;id=fqVYcFtBr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI’s ChatGPT atlas browser architecture</strong>: OpenAI’s ChatGPT Atlas browser uses a unique architecture that runs Chromium as a separate process, enabling instant startup, responsiveness with hundreds of tabs, and rich SwiftUI animations through layer hosting and IPC. <a href="https://app.daily.dev/posts/TsFQIm4Vj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WebMCP standard for AI agent interaction in browsers</strong>: Chrome is previewing WebMCP, a proposed web standard allowing websites to expose structured tools directly to in-browser AI agents via explicit function calls, replacing brittle DOM scraping with declarative HTML annotations. <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>Datadog and github copilot launch AI-powered testing tools</strong>: Datadog launched Code Coverage with AI-generated tests for uncovered code paths, while GitHub Copilot Testing for .NET is now generally available, offering AI-powered unit test generation with natural language prompting. <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>
<h2>Bullshit police</h2>
<ul>
<li><strong><a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a> releases glm-5, claims parity with proprietary models</strong>: <a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a> just dropped GLM-5, an open-source model they claim matches Claude Opus 4.6 and GPT 5.3 Codex for long-horizon coding tasks. It’s 744B parameters, 40B active, and trained on 28.5T tokens. Sure, it’s open-source, but these claims of matching proprietary models are getting a bit repetitive, aren’t they? Let’s see if it actually delivers or if it’s just another flavor of the month. <a href="https://app.daily.dev/posts/bOspGaOnW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will make you work longer, not smarter</strong>: A Harvard Business Review study found that generative AI makes tech workers take on more tasks, reduce breaks, and work longer hours, leading to increased cognitive fatigue and burnout. So much for AI making our lives easier, huh? Turns out, it just gives us more rope to hang ourselves with, voluntarily of course. <a href="https://app.daily.dev/posts/mEvZkBSFL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are the new hotness, but still need humans to clean up</strong>: Google Cloud launched GEAR to help build and deploy AI agents, while Microsoft Foundry is now an ‘end-to-end enterprise AI agent platform.’ Everyone’s pushing these autonomous agents, promising 30-50% efficiency gains. Yet, Deloitte says only 25% of enterprises productionize 40% of their AI pilots. So, we’re still building fancy tools that mostly stay in pilot purgatory, and when they do escape, they need ‘human-in-the-loop’ workflows. Progress! <a href="https://app.daily.dev/feed-by-ids?id=9C4SjPZli&amp;id=gwxHevbDX&amp;id=o13HbD6JR&amp;id=r88T1MoRe&amp;id=vS730ONZY&amp;id=dz3njIXXE&amp;id=it7NMfnEb&amp;id=laJOXndNn&amp;id=tEzN0YnW4&amp;id=wbweznlAO" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
