<h2>Must know</h2>
<ul>
<li><strong>CISA warns of active exploitation in VMware vCenter</strong>: CISA has confirmed active exploitation of CVE-2024-37079, a critical heap overflow vulnerability in VMware vCenter Server with a CVSS score of 9.8. This flaw allows remote code execution and can be chained to gain root access to ESXi. Federal agencies are mandated to patch by February 13, 2026, highlighting the severity and widespread risk, as reported by Security Boulevard. <a href="https://app.daily.dev/posts/sI8icXzRd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are a security nightmare for IAM</strong>: AI agents are fundamentally breaking traditional Identity and Access Management (IAM) models because they compile broad, persistent permissions without clear ownership or accountability. This creates significant authorization risks and pathways for bypasses, especially with ‘organizational agents’ shared across teams. New security protocols are urgently needed to track relationships between users, agents, and systems, as current models are ill-equipped to handle autonomous AI actions. <a href="https://app.daily.dev/posts/nUFRFzE1V" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft and IBM push quantum development forward</strong>: Microsoft has upgraded its Quantum Development Kit with AI-assisted programming, integrating GitHub Copilot and Visual Studio Code to lower barriers for quantum development. Meanwhile, IBM Quantum Platform continues to focus on cloud-based access to real quantum hardware, emphasizing Qiskit for programming and advanced error mitigation strategies. Both aim to bridge the gap between current error-prone qubits and future fault-tolerant systems. <a href="https://app.daily.dev/feed-by-ids?id=7ZfY82yLW&amp;id=MiiMPahSr" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Google Cloud opens public datasets via Apache Iceberg</strong>: Google Cloud now offers public datasets, like the NYC Taxi dataset, through the Apache Iceberg REST Catalog on BigLake, allowing read-only access by any compute engine supporting the Iceberg protocol without infrastructure setup. <a href="https://app.daily.dev/posts/tn371FITl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oracle’s Orasort patent expires, open source adoption expected</strong>: Oracle’s ‘Orasort’ sorting algorithm patent expired, making the 5x faster algorithm available for open-source use, with immediate interest from MySQL and PostgreSQL communities for database performance enhancements. <a href="https://app.daily.dev/posts/aheyA2T6F" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Compyle slashes cloud terminal latency to 14ms</strong>: Compyle drastically reduced cloud development environment terminal latency from 200ms to 14ms by implementing a warm pool, removing the 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>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>DXVK-NVAPI 0.9.1 improves DLSS and fixes game crashes</strong>: DXVK-NVAPI version 0.9.1 introduces environment variables for manual DLSS scaling ratio overrides 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>Microsoft Aspire expands to support multiple languages</strong>: Microsoft’s Aspire platform now supports JavaScript, TypeScript, Python, and Java as first-class languages for distributed application development, broadening its appeal beyond .NET users, especially for AI/ML and web development. <a href="https://app.daily.dev/posts/gRAkh5lrI" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI coding agents: the new energy hog in your data center</strong>: Apparently, your AI coding assistant, like Claude Code, is a power guzzler. Tech Lead Digest reports a single session can use as much electricity as running a dishwasher, with daily usage comparable to a refrigerator. So, while you’re busy ‘optimizing’ your code with AI, you’re also optimizing your electricity bill. Who knew developer productivity came with such a carbon footprint? <a href="https://app.daily.dev/posts/ex1Cm3GIg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents and ‘code mode’: solving context window exhaustion with more complexity</strong>: We’re told ‘code mode’ for AI agents, like Port of Context, will reduce token usage by 98.7% by letting AI write and execute sandboxed code. Because the solution to LLMs struggling with context is obviously to give them more rope to hang themselves with, by letting them write and run their own code. What could possibly go wrong with autonomous agents executing code in a Deno runtime? Nothing, I’m sure. <a href="https://app.daily.dev/feed-by-ids?id=F3rJpKxUE&amp;id=mcXaLHnCP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Coi language: beating React and Vue with… more claims</strong>: A new language called Coi claims to ‘beat’ React and Vue in UI performance by compiling to WASM and using a shared memory architecture. Their benchmarks show 100 FPS canvas rendering versus Emscripten’s 40 FPS. It’s always great to see a new contender, especially one making bold claims without any independent verification. I’m sure this isn’t just another ‘Show HN’ hype cycle, right? <a href="https://app.daily.dev/posts/iOGOTjzat" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
