<h2>Must know</h2>
<ul>
<li><strong>Salesforce shifts Heroku to sustaining engineering</strong>: Salesforce has moved its Platform-as-a-Service (PaaS) Heroku into a sustaining engineering phase, halting new enterprise contract sales. This move is widely interpreted as a precursor to phasing out the platform, which has struggled with rising costs and competition from alternatives like Render and Vercel. Existing customers are advised to treat Heroku as legacy infrastructure and plan migrations, as reported by various analysts. <a href="https://app.daily.dev/posts/13PGZg8yI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents develop C compiler for Linux kernel</strong>: Anthropic researchers used 16 Claude Opus 4.6 AI agents to autonomously build a Rust-based C compiler capable of compiling the Linux kernel. The experiment, costing $20,000 in API calls over two weeks, generated 100,000 lines of code and successfully built Linux 6.9 for multiple architectures. While the generated code is not production-ready due to inefficiency and subpar Rust quality, it demonstrates significant autonomous AI development capabilities, as detailed in the research. <a href="https://app.daily.dev/posts/qhPl6UlEO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>EU accuses Meta of WhatsApp AI antitrust violation</strong>: The European Commission has formally accused Meta of violating EU competition rules by blocking third-party AI assistants like ChatGPT and Microsoft Copilot from WhatsApp. Following policy changes in October 2025, Meta AI became the sole AI assistant allowed on the platform, effective January 2026. The EU regulator is considering emergency measures to restore competitor access, arguing Meta’s restrictions could seriously harm the AI assistant market. <a href="https://app.daily.dev/posts/w9RLQf8YH" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI’s dual impact on software development</strong>: AI is causing anxiety with ‘10x productivity’ expectations, shifting developer roles, and threatening open-source sustainability through ‘vibe coding’ and overwhelming maintainers with low-quality AI-generated contributions. <a href="https://app.daily.dev/feed-by-ids?id=0CiZDAswQ&amp;id=0hMgfxxrs&amp;id=7G6kpg12b&amp;id=8yRieynOP&amp;id=KCC82Mw6D&amp;id=d4YFQW1JT&amp;id=eD5zl44R8&amp;id=gVU0l8bpQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel permanently adopts Rust</strong>: Linux 7.0 officially makes Rust a permanent part of the kernel project, signaling a long-term commitment and encouraging further investment in Rust development and training within the community. <a href="https://app.daily.dev/posts/k0qXAqGCH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents influence programming language design</strong>: AI agents are driving new programming language design considerations, favoring explicit typing, avoiding whitespace, supporting effect markers, and prioritizing deterministic testing and clear tooling for optimal agent workflows. <a href="https://app.daily.dev/posts/PKgkpakDW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cluster API v1.12 enhances Kubernetes lifecycle management</strong>: Cluster API v1.12.0 introduces in-place updates for modifying existing machines without recreation and chained upgrades for skipping multiple Kubernetes minor versions in a single operation, simplifying cluster lifecycle management. <a href="https://app.daily.dev/posts/SwlgWxJBR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Valkey performance enhancements post-Redis fork</strong>: Valkey, a Redis fork, is undergoing significant performance improvements in versions 8 and 9, including a redesigned hash table, reduced memory overhead, and improved throughput while maintaining Redis 7.2 compatibility. <a href="https://app.daily.dev/posts/BJPWCn9Zl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s micro frontend strategy and TypeScript’s future</strong>: Google’s micro frontend approach emphasizes complete isolation with Protocol Buffers for communication, while TypeScript 6.0 will be the final version written in TypeScript before a planned rewrite in Go. <a href="https://app.daily.dev/posts/OBdZWzGyV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Modern CSS is a platform, not a styling language</strong>: Modern CSS has evolved into a declarative UI runtime with features like container queries, :has(), and cascade layers, enabling complex responsive behaviors and reducing reliance on JavaScript for UI logic. <a href="https://app.daily.dev/posts/F72fOvEL9" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic Claude co-work plugin release and tech stock market concerns</strong>: Anthropic released Claude Co-work with industry-specific plugins, and some sources are claiming this ‘wiped out $400 billion in tech stocks.’ Because, you know, a single AI plugin release is totally responsible for market fluctuations. It’s not like there are a million other factors at play, but sure, let’s blame the new AI toy. <a href="https://app.daily.dev/posts/QnkkrfiEd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI’s impact on cognitive skills: developer concerns</strong>: Two unnamed developers are worried that AI tools like ChatGPT are fostering ‘lazy thinking’ because they’re personally offloading cognitive tasks to AI. It’s almost as if tools designed to automate tasks might, gasp, automate tasks. Maybe the problem isn’t the AI, but the human’s inability to, you know, <em>think</em> when the crutch is removed? <a href="https://app.daily.dev/posts/X3beZyrKi" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated compiler performance analysis</strong>: Claude’s C Compiler (CCC) successfully compiled the Linux kernel, but the binaries ran 737 to 158,000 times slower than GCC. It also used 5.9 times more memory and produced 2.7 to 3 times more code bloat. So, AI can build complex software, but it’s about as efficient as a snail trying to run a marathon. Progress! <a href="https://app.daily.dev/posts/9jEdtbewr" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
