<h2>Must know</h2>
<ul>
<li><strong>Ladybird browser pivots to rust with AI assistance</strong>: The Ladybird browser project is ditching Swift for Rust, primarily for memory safety. Lead developer Andreas Kling used AI agents like Claude Code and Codex to translate 25,000 lines of C++ to Rust in just two weeks, a task estimated to take months manually. This human-directed AI workflow resulted in byte-for-byte identical output and zero test regressions, demonstrating AI’s potential for accelerating large-scale refactoring, as reported by Changelog News. <a href="https://app.daily.dev/feed-by-ids?id=bavH6r0ZC&amp;id=1QrlHS5ue" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic accuses chinese labs of industrial-scale model distillation</strong>: Anthropic claims Chinese AI labs DeepSeek, Moonshot AI, and MiniMax created over 24,000 fraudulent accounts to distill its Claude models, generating 16 million exchanges to extract capabilities like agentic reasoning and coding. Anthropic frames this as a national security threat, arguing distilled models lack safety guardrails and could be misused. This accusation comes amidst ongoing US debates on AI chip export controls to China, with Anthropic using these incidents to push for stricter restrictions. <a href="https://app.daily.dev/posts/YcZmCF704" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-augmented FortiGate firewall compromises reported by AWS</strong>: AWS detected a financially motivated, Russian-speaking threat actor compromising over 600 FortiGate firewall instances across 55+ countries. This unsophisticated group leveraged generative AI tools for reconnaissance, exploitation instructions, script writing, and operational notes, operating at a scale previously requiring larger, skilled teams. The attacks exploited exposed management interfaces, weak credentials, and lack of MFA, highlighting that basic security hygiene would have prevented most of the activity. <a href="https://app.daily.dev/posts/azBzuZltn" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Uber open-sources uForwarder kafka consumer proxy</strong>: Uber has open-sourced uForwarder, a push-based Kafka consumer proxy designed to handle trillions of messages daily, centralizing offset management and simplifying consumer logic for microservices. <a href="https://app.daily.dev/posts/NOSp6Xv2e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub introduces Agent HQ for unified AI coding agents</strong>: GitHub announced Agent HQ, a unified platform for coding agents like Claude, Codex, and GitHub Copilot, allowing users to run and compare multiple frontier agents simultaneously with a single subscription. <a href="https://app.daily.dev/posts/CxN7d4Xat" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cohere releases aya tiny multilingual models</strong>: Cohere launched the Aya Tiny model family, small (~3.3B parameter) multilingual language models supporting over 70 languages, designed for mobile devices and underserved linguistic communities. <a href="https://app.daily.dev/posts/QusoaclzE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI releases gpt-realtime-1.5 model</strong>: OpenAI released gpt-realtime-1.5 via its Realtime API, enhancing voice workflows with more reliable instruction following, improved tool calling, and better multilingual accuracy for low-latency speech-to-speech interactions. <a href="https://app.daily.dev/posts/6I1BavGOr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 148 release with AI controls</strong>: Firefox 148 introduces an ‘AI kill switch’ allowing users to disable all AI-powered features or selectively block individual ones, alongside web platform updates like Trusted Types API and Service Worker support for WebGPU. <a href="https://app.daily.dev/posts/UNAL37VrS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ZeroDayRAT mobile spyware emerges, bypassing MFA</strong>: ZeroDayRAT, a new mobile spyware sold on Telegram, targets Android 5-16 and iOS up to 26, offering capabilities like keylogging, SMS control (for MFA bypass), and credential theft, commoditizing nation-state-level surveillance. <a href="https://app.daily.dev/posts/Q0yMWpjwU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Starkiller phishing kit bypasses MFA</strong>: Starkiller, a Phishing-as-a-Service toolkit, bypasses MFA and traditional detection by live-proxying login pages and capturing session tokens, lowering the technical barrier for attackers. <a href="https://app.daily.dev/posts/9TGsm3aiu" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Google DeepMind claims image generation breakthrough</strong>: Apparently, Google DeepMind ‘just solved’ one of the ‘dirtiest problems’ in image generation with an ‘embarrassingly elegant’ fix, according to some random Twitter user. No actual details, no official announcement, just pure, unadulterated hype from @rryssf_. Sounds like another groundbreaking discovery that will remain a mystery to everyone outside of their echo chamber. <a href="https://app.daily.dev/posts/fzJvtM3B4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI updates AI coding evaluation standard</strong>: OpenAI, in a move that absolutely no one saw coming, has decided their previous benchmark, SWE-bench Verified, is ‘saturated’ and ‘contaminated.’ So now, they recommend SWE-bench Pro. Because when your AI models get too good for your own tests, you just invent a new, ‘more rigorous’ one, right? All according to OpenAI, of course. <a href="https://app.daily.dev/posts/oZGyDrjcX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude code security launch and market reaction</strong>: Anthropic launched Claude Code Security, an AI tool that supposedly finds vulnerabilities, causing cybersecurity stocks to tank. Analysts called the market reaction ‘illogical,’ probably because they realize an AI that ‘understands component interactions’ isn’t going to replace actual security professionals. It’s a ‘limited research preview’ for ‘enterprise and team customers,’ so let’s not get ahead of ourselves with the AI savior narrative. <a href="https://app.daily.dev/posts/bgcFYmVK1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
