<h2>Must know</h2>
<ul>
<li><strong>PackageGate zero-days expose npm and yarn to worm attacks</strong>: Six zero-day vulnerabilities, dubbed PackageGate, are affecting JavaScript package managers like npm and yarn, allowing attackers to bypass security controls designed to prevent worm attacks. While pnpm, vlt, and Bun have already patched these flaws, npm and yarn have not, with npm claiming the behavior is ‘by design.’ Researcher Oren Yomtov strongly advises developers to switch to patched package managers to avoid potential malicious payload delivery. This is a critical vulnerability for JavaScript developers, as it enables worm attacks despite existing mitigations. <a href="https://app.daily.dev/posts/klUkuFZvK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Office zero-day forces emergency patch</strong>: Microsoft has issued an emergency out-of-band patch for a critical zero-day vulnerability (CVE-2026-21509) in Microsoft Office, actively exploited in the wild. This flaw allows attackers to bypass security protections using malicious Office files, primarily targeting legacy COM and OLE components. While newer Office versions and Microsoft 365 Apps are patched, Office 2016 and 2019 users must rely on manual registry workarounds. CISA has mandated all federal agencies patch this by February 16, 2026, highlighting the urgency of this threat. <a href="https://app.daily.dev/posts/qC937DXHe" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI use accelerates decline of public knowledge platforms</strong>: AI models like ChatGPT and Claude are increasingly replacing public knowledge platforms such as Stack Overflow and Wikipedia for problem-solving, leading to a reported 78% drop in Stack Overflow traffic. This shift risks a ‘model collapse’ where AI systems degrade by training on increasingly AI-generated content, detaching from human expertise. The articles suggest this fragments open collaboration and collective learning, favoring private, immediate solutions over communal knowledge, and raising concerns about corporate AI monopolies consolidating power. <a href="https://app.daily.dev/posts/3Qlqmk7FI" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Teleport unveils agentic identity framework for AI security</strong>: Teleport has launched a new Agentic Identity Framework that uses cryptographic authentication for AI agents, moving away from static credentials to reduce risks like credential theft and prompt injection attacks. <a href="https://app.daily.dev/posts/CI1C8FH63" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents reshape developer roles and enterprise operations</strong>: AI is shifting developer roles from coding to product architecture and outcome engineering, with enterprises integrating AI agents as core system features, necessitating new infrastructure for orchestration and fault tolerance, and robust governance to prevent ‘agent sprawl.’ <a href="https://app.daily.dev/feed-by-ids?id=b2fEANN64&amp;id=yBxtGCVqr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare launches serverless post-quantum Matrix homeserver</strong>: Cloudflare has ported a full Matrix homeserver, Tuwunel, to its Workers platform, offering a serverless architecture with X25519MLKEM768 post-quantum cryptography by default, aiming for sub-50ms global latency and minimal operational overhead. <a href="https://app.daily.dev/posts/vURH9h7xw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>CISA publishes federal guidance on post-quantum cryptography adoption</strong>: CISA has released guidance mapping NIST-standardized post-quantum cryptography algorithms to enterprise hardware and software, directing federal agencies to procure PQC-capable products where solutions are available, highlighting adoption gaps in digital signatures and authentication. <a href="https://app.daily.dev/posts/5Lj7cpvbX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>D-Wave expands quantum computing and secures major enterprise contract</strong>: D-Wave announced significant progress in its quantum computing roadmap, including 314% growth in Advantage2 annealing system usage, integration of ML into its Stride solver, and a $10 million QCaaS contract with a Fortune 100 company, aiming to launch its first gate-model quantum system in 2026. <a href="https://app.daily.dev/posts/B9UXF3b8Z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WhatsApp announces strict account settings for high-risk users</strong>: WhatsApp launched ‘Strict Account Settings’ for high-risk users like journalists, offering automatic media blocking from unknown senders, silenced calls, two-step verification, and restricted profile visibility, following a lawsuit against Meta regarding privacy claims. <a href="https://app.daily.dev/posts/0u5TXEYDl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pinterest workforce reduction and AI strategy shift</strong>: Pinterest is laying off nearly 15% of its workforce to reallocate resources towards AI-focused roles and open-source AI models, aiming to develop AI-powered products like Pinterest Assistant and personalized boards. <a href="https://app.daily.dev/posts/yPKU3lKW1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI coding agents: still more hype than help?</strong>: OpenAI’s own technical breakdown of Codex CLI admits its AI coding agents are ‘brittle and prone to errors’ outside their training data, requiring human oversight for production. Meanwhile, an experiment with AI agents in Swift projects found generated code often violates language idioms and introduces ‘unnecessary complexity.’ So, while they’re great for boilerplate, don’t expect them to write your next clean, maintainable codebase without a lot of hand-holding. It seems the ‘AI as a junior developer’ metaphor is still holding strong, and not in a good way. <a href="https://app.daily.dev/feed-by-ids?id=qZX3wHxMK&amp;id=uilUGnhp4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare’s Matrix homeserver: a quantum leap in… incompleteness?</strong>: Cloudflare proudly announced a Matrix homeserver on its Workers platform, touting serverless architecture and post-quantum crypto. However, a Matrix developer quickly pointed out it lacks critical security features like authorization, is riddled with ‘TODO’ comments, and might cost more than a standard VPS. The critique even suggested it might be ‘partially AI-generated.’ So, a ‘complete’ homeserver that’s incomplete, insecure, and potentially more expensive? Sounds like a classic case of marketing getting ahead of reality. <a href="https://app.daily.dev/feed-by-ids?id=vURH9h7xw&amp;id=93HDTbjA4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ricursive Intelligence: the $4 billion startup that does… something with AI chips?</strong>: Ricursive Intelligence, founded by ex-Google researchers, just snagged $300 million at a $4 billion valuation, two months after launch. Their grand plan? An AI system to design and ‘automatically improve AI chips.’ No products, no public statements from founders, and zero independent verification of their ‘AlphaChip’ claims. It’s another one of those ‘trust us, it’s AI, and it’s worth billions’ stories. Because apparently, in the AI world, a good origin story and a vague promise are all you need for a Series A. <a href="https://app.daily.dev/posts/pqGutcjN8" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
