<h2>Must know</h2>
<ul>
<li><strong>Google engineers indicted for stealing trade secrets and transferring to Iran</strong>: Two former Google engineers and sisters, Samaneh and Soroor Ghandali, along with an alleged accomplice, have been federally indicted for stealing trade secrets related to processor security, cryptography, and Snapdragon SoC hardware architecture. Prosecutors allege the data was transferred to Iran using methods like Telegram and photographing screens to evade detection. This highlights significant insider threat risks and the lengths bad actors will go to exfiltrate sensitive intellectual property, as reported by multiple sources. <a href="https://app.daily.dev/posts/hfUKRpTfu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GGML and llama.cpp join Hugging Face for local AI growth</strong>: Georgi Gerganov and the GGML team, creators of the popular <code>llama.cpp</code> project, are joining Hugging Face. This collaboration aims to ensure the long-term sustainability and growth of local AI inference, with Hugging Face providing resources while the GGML team retains full technical autonomy. The goal is to make open-source AI inference accessible on consumer hardware, positioning local inference as a competitive alternative to cloud solutions. <a href="https://app.daily.dev/posts/ONSshEfvD" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Red Hat launches enterprise Podman Desktop</strong>: Red Hat has released an enterprise-supported build of Podman Desktop in technical preview, targeting corporate developers who need Docker-compatible container workflows. This offering comes with vendor-backed SLAs, security fixes, and policy enforcement, integrating deeply with OpenShift and RHEL. It aims to compete with Docker Desktop by emphasizing enterprise features and open-source credentials. <a href="https://app.daily.dev/posts/jbEPljaPx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Go 1.26 introduces rewritten <code>go fix</code> tool</strong>: Go 1.26 features a rewritten <code>go fix</code> tool with 24 transformations for code modernization, including <code>interface{}</code> to <code>any</code> and <code>strings.Cut</code> adoption, improving code quality and maintainability. <a href="https://app.daily.dev/posts/0hCCBK1Iz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hardware industry shifts to AI data centers, causing consumer shortages</strong>: A structural shift towards AI data centers and hyperscalers is causing severe shortages and price increases (15-95%) for consumer DRAM, NAND, and HDDs, with Micron exiting the consumer memory market and Western Digital/Kioxia selling out 2026 production to enterprises. <a href="https://app.daily.dev/posts/2aSZDslGW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Erlang/OTP patch releases fix critical ETS bug and security vulnerability</strong>: Erlang/OTP 27.3.4.8 and 26.2.5.17 patch releases address a critical ETS bug where <code>ets:update_counter/4</code> and <code>ets:update_element/4</code> could crash the ERTS, along with a path traversal vulnerability in the <code>tftp</code> module. <a href="https://app.daily.dev/feed-by-ids?id=TXXfoJq3M&amp;id=iDNopOiir" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI launches enterprise AI agent platform “Frontier”</strong>: OpenAI introduced “Frontier,” an enterprise-focused platform for building, deploying, and managing AI agents at scale, offering tooling for the full development lifecycle and integration with existing systems. <a href="https://app.daily.dev/posts/NGz762AB9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Halodoc migrates Angular SSR monolith to Angular 20</strong>: Halodoc successfully migrated its Angular SSR monolith to Angular 20, resolving 12 challenges including switching from Webpack to Vite+esbuild for 56% faster builds and upgrading to Express 5. <a href="https://app.daily.dev/posts/TzZRtuZ8M" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloud Hypervisor 51 improves performance and QCOW2 v3 support</strong>: Cloud Hypervisor 51, a Rust-based Virtual Machine Monitor, has been released with performance enhancements and improved QCOW2 v3 support, focusing on secure cloud computing. <a href="https://app.daily.dev/posts/aAFeVuDlX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Intel hiring more Linux developers for GPU drivers and gaming</strong>: Intel is actively hiring Linux software developers, specifically for GPU driver development and the Linux gaming stack, with a focus on improving gaming compatibility, including Valve’s Proton. <a href="https://app.daily.dev/posts/lEJij6iSi" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents need human guidance to perform better</strong>: New research on SkillsBench, a benchmark for agentic AI, found that AI agents perform significantly better (16.2 percentage points) with curated, human-provided skills. Apparently, these ‘autonomous’ agents can’t figure out how to improve on their own, or even generate useful skills. So, after all the hype, it turns out they still need us to hold their digital hands. Shocking. <a href="https://app.daily.dev/posts/c3BmZNVYL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Claude Code desktop gets new features</strong>: Anthropic’s Claude Code on desktop now lets you preview running apps, review code, and handle CI failures and pull requests in the background. Because nothing says ‘cutting-edge AI’ like features that have been standard in IDEs for years, now repackaged with a chatbot. It’s almost like they’re trying to catch up to basic developer tooling, but with more LLM-flavored marketing. <a href="https://app.daily.dev/posts/1pjLIO04t" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent development is transforming software, but security is a top concern</strong>: Reports claim 90% of software professionals use AI agents daily and 41% of global code is AI-generated, yet 96% don’t trust AI-generated code without manual review, spending 24% of their week auditing it. So, we’re supposedly transforming software development with AI, but everyone’s still spending a quarter of their time cleaning up after it. Sounds less like a revolution and more like a ‘toil swap’ with extra steps and a side of security nightmares. <a href="https://app.daily.dev/posts/0RZnFaVdx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
