<h2>Must know</h2>
<ul>
<li><strong>AI agents are getting serious about open source reputation farming</strong>: An AI agent named Kai Gritun submitted over 100 pull requests across nearly 100 open-source repositories in just days, a tactic security experts are calling ‘reputation farming.’ This rapid accumulation of influence in critical projects could enable future supply chain attacks, shifting the attack surface from code to governance processes. This is a new twist on the XZ-utils backdoor, which took years to build trust, but now AI can do it in a fraction of the time. <a href="https://app.daily.dev/feed-by-ids?id=D9JdnEIyW&amp;id=QHtbtBXdQ&amp;id=UNO27M8rA&amp;id=VGEsVlZIo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Chrome zero-day exploited in the wild</strong>: Google just pushed an emergency Chrome update to patch CVE-2026-2441, a high-severity use-after-free vulnerability in CSS handling that’s already being actively exploited. This marks Chrome’s first zero-day of 2026, following eight last year. If you haven’t updated, do it now, unless you enjoy arbitrary code execution via malicious webpages. <a href="https://app.daily.dev/feed-by-ids?id=L6DW92msi&amp;id=MVqzUbkKx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is changing software development and developer roles</strong>: AI coding tools are fundamentally shifting developer roles from pure coders to orchestrators and validators. Poor documentation is now a critical risk, as AI relies on it and will ‘hallucinate’ without it. Over-reliance on AI could also lead to ‘cognitive atrophy’ in programmers, diminishing problem-solving skills. Meanwhile, multi-agent systems like Gas Town are emerging, coordinating dozens of AI agents in parallel on a single codebase, albeit at a hefty $100/hour. <a href="https://app.daily.dev/feed-by-ids?id=2XGWtlWnL&amp;id=JV1i29nbu&amp;id=WYcCi2eVO&amp;id=Wvk48DEZ7&amp;id=dVvuHhcvo&amp;id=oZ8mOxinB&amp;id=qtVi65Njq" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>OpenAI hires OpenClaw creator for next-gen AI agents</strong>: Peter Steinberger, creator of the popular open-source AI agent OpenClaw, has joined OpenAI to lead development of next-generation personal AI agents, with OpenClaw transitioning to an open-source foundation supported by OpenAI. <a href="https://app.daily.dev/feed-by-ids?id=2BTVSGYkh&amp;id=J7xnq8nba&amp;id=QkDNtEDZb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>PostgreSQL scales to 800 million users at OpenAI</strong>: OpenAI scaled PostgreSQL to handle millions of queries per second for 800 million ChatGPT users using a single-primary architecture with read replicas, focusing on minimizing primary load and optimizing queries. <a href="https://app.daily.dev/posts/0QBTpoQPy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Chrome 145 features for developers</strong>: Chrome 145 enhances DevTools with soft navigation visibility, precise profiler timings, and AI assistance, while WebGPU gains <code>subgroup_uniformity</code> and <code>mapSync()</code> for improved performance. <a href="https://app.daily.dev/feed-by-ids?id=AK3a3hE0z&amp;id=EWohDzjzG&amp;id=bvXFp0lun" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Java gets null safety and C-like performance</strong>: Java is introducing null safety via Project Valhalla, using null-restricted value types and aiming for C-like performance by eliminating pointer chasing, all while maintaining backward compatibility. <a href="https://app.daily.dev/posts/YYmRfNhUM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Modern CSS handles complex UI tasks without JavaScript</strong>: Modern CSS has evolved to manage complex UI tasks like custom selects, anchor positioning, and scroll-driven animations, reducing the need for JavaScript for most presentation and interaction needs. <a href="https://app.daily.dev/feed-by-ids?id=1TO1NBQ6P&amp;id=E3ttiNVjr&amp;id=G7unzK9jJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Agoda open-sources API agent for natural language queries</strong>: Agoda released API Agent, an open-source MCP server that connects to any REST or GraphQL API, introspects schemas, and generates queries from natural language input without requiring code. <a href="https://app.daily.dev/posts/VhIIJvhkg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Small language models for local deployment</strong>: Seven small language models (SLMs) from 1B to 9B parameters are now available for local execution on consumer laptops, offering specialized performance for tasks like RAG, general purpose, and code generation. <a href="https://app.daily.dev/posts/Uggt1hDI5" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>NVIDIA’s Blackwell Ultra promises 50x performance for agentic AI</strong>: NVIDIA’s Blackwell Ultra platform is supposedly delivering up to 50x higher throughput and 35x lower cost for agentic AI workloads. Because, you know, we totally needed another ‘revolutionary’ hardware claim to justify those GPU prices. I’m sure all those agentic AI workloads are just waiting for this magical hardware to unlock their true potential. <a href="https://app.daily.dev/posts/XLkxKsrjj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated malware exploits honeypots, security experts warn</strong>: Attackers are now using LLM-generated malware to exploit vulnerabilities in Docker honeypots, leading ‘security experts’ to warn of a ‘shift in the threat landscape.’ Apparently, AI tools are enabling ‘low-skilled operators’ to build exploitation frameworks. So, the bad guys are getting smarter, or just lazier, and now we have to worry about AI writing bad code too. <a href="https://app.daily.dev/posts/hzvvYMQze" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is making open source a spam-filled nightmare</strong>: The curl project dropped bug bounties because AI-generated spam drowned out useful vulnerability reports, and open-source maintainers are getting overwhelmed by low-quality AI-generated pull requests. GitHub is even letting repos disable PRs entirely. So much for AI ‘democratizing’ development; it’s just democratizing garbage. <a href="https://app.daily.dev/feed-by-ids?id=8eYbRwAri&amp;id=p4VqhMMFm" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
