<h2>Must know</h2>
<ul>
<li><strong>CISA warns of active exploitation of VMware vCenter flaw</strong>: CISA has confirmed active exploitation of VMware vCenter Server vulnerability CVE-2024-37079, a critical heap overflow flaw with a CVSS score of 9.8. This vulnerability allows remote code execution and can be combined with other exploits to gain root access to ESXi. Federal agencies are mandated to patch affected systems by February 13, 2026, highlighting the severe risk this poses to critical infrastructure. <a href="https://app.daily.dev/posts/sI8icXzRd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents pose significant IAM security risks</strong>: AI agents are fundamentally incompatible with traditional Identity and Access Management (IAM) systems, creating urgent security risks. Unlike human users, agents often compile broad, persistent permissions without clear ownership or accountability, leading to potential authorization bypasses. Experts warn that ‘organizational agents’ shared across teams are particularly dangerous, necessitating new security protocols to track relationships between users, agents, and systems. <a href="https://app.daily.dev/posts/nUFRFzE1V" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oracle’s Orasort patent expires, opening door for open source adoption</strong>: Oracle’s patent on the ‘Orasort’ sorting algorithm expired on January 24, 2026, making it available for open-source use. This algorithm, invented by Mark Callaghan and implemented in Oracle 10gR2, offers a 5x performance improvement over older methods by combining common prefix skipping, adaptive quicksort/radix sort switching, and substring key caching. Its availability is expected to significantly enhance database sorting performance for projects like MySQL and PostgreSQL. <a href="https://app.daily.dev/posts/aheyA2T6F" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Microsoft and IBM push quantum development tools</strong>: Microsoft upgraded its Quantum Development Kit with AI-assisted programming and GitHub Copilot integration, while IBM continues to focus on its cloud-based Quantum Platform with Qiskit for real hardware experimentation. <a href="https://app.daily.dev/feed-by-ids?id=7ZfY82yLW&amp;id=MiiMPahSr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GNOME AI assistant Newelle gets local LLM support</strong>: Newelle, the AI assistant for GNOME, now supports local LLM execution via llama.cpp and includes a system command execution tool, offering users a choice between cloud-based and on-device AI inferencing. <a href="https://app.daily.dev/posts/Au9UNs8LJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Cloud offers public datasets via Apache Iceberg REST Catalog</strong>: Google Cloud has made public datasets, like the NYC Taxi dataset, accessible as read-only Iceberg tables through the Apache Iceberg REST Catalog on BigLake, allowing easy querying without infrastructure setup. <a href="https://app.daily.dev/posts/tn371FITl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Compyle drastically reduces cloud terminal latency</strong>: Compyle reduced cloud development environment terminal latency from 200ms to 14ms by implementing a warm pool, removing a socket server middleman, and deploying multi-regional sandbox pools closer to users. <a href="https://app.daily.dev/posts/2R6Ajv0Sw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>DXVK-NVAPI 0.9.1 adds DLSS ratio override and fixes</strong>: DXVK-NVAPI version 0.9.1 introduces environment variables for manually overriding DLSS scaling ratios and fixes startup crashes in several games, including ‘Middle-earth: Shadow of Mordor’ and ‘The Last of Us Part 2’. <a href="https://app.daily.dev/posts/6hOCF9MBP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Skip tool open-sourced for cross-platform mobile development</strong>: Skip, a tool that converts Swift code for native iOS and Android operation, has been fully open-sourced, removing all paywalls and shifting to community funding via GitHub Sponsors. <a href="https://app.daily.dev/posts/7snXi8ahF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Servo 0.0.4 released with multi-window support and security improvements</strong>: Servo 0.0.4 introduces multi-window support, HTTP proxy configuration, and enables system root certificates by default, enhancing security and compatibility for the browser engine. <a href="https://app.daily.dev/posts/72B89BNSg" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents are apparently too dumb for novel solutions</strong>: Turns out, AI coding agents struggle with anything truly ‘novel’ because they’re just regurgitating patterns from their training data, according to Rubyland. If you want an LLM to do something outside the box, you have to trick it by calling a checker piece a UFO. So much for groundbreaking innovation, just more fancy pattern matching. <a href="https://app.daily.dev/posts/gov03HPhm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will make you an SRE, because writing code is too easy now</strong>: Two identical articles claim AI makes writing code so trivial that software engineers’ new core value is operational excellence and reliability. Apparently, building a working demo is ‘straightforward’ with AI, but keeping it running is the real challenge. So, instead of being creative problem-solvers, we’re all just going to be glorified SREs, cleaning up after AI’s ‘straightforward’ code. Thanks, AI, for making our jobs ‘easier’. <a href="https://app.daily.dev/posts/otuC5YedQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Coi language claims to beat React/Vue, because benchmarks are always trustworthy</strong>: A new component-based language called Coi is here, claiming ‘performance superiority’ over React and Vue by compiling to WASM. Their own benchmarks show 100 FPS canvas rendering versus Emscripten’s 40 FPS. Because self-reported benchmarks from a ‘Show HN’ announcement are definitely the gold standard for objective technical assessment. I’m sure there are absolutely no caveats or independent validations needed here. <a href="https://app.daily.dev/posts/iOGOTjzat" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
