<h2>Must know</h2>
<ul>
<li><strong>AI agent security vulnerabilities are a growing concern</strong>: A new report from Snyk’s ToxicSkills research found that 13.4% of AI agent skills contain critical security issues, with 76 confirmed malicious payloads designed for credential theft and data exfiltration. This highlights a significant supply chain risk for enterprises adopting AI agents. Separately, Operant AI launched Agent Protector, a real-time security solution for autonomous AI agents, offering inline protection against privilege escalation and data exfiltration, emphasizing that platform providers should be responsible for agent security, not customers. <a href="https://app.daily.dev/feed-by-ids?id=0BdgmYNfg&amp;id=0e28QTlYb&amp;id=2vNCw5dB4&amp;id=dL9ys7HAk&amp;id=the2EenMO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft appoints engineering quality czar</strong>: Microsoft CEO Satya Nadella appointed Charlie Bell as the new engineering quality czar, reporting directly to him. This move aims to address ongoing issues like Azure outages, Windows patch problems, and low Copilot adoption rates, signaling a renewed focus on product reliability and user experience across Microsoft’s offerings. <a href="https://app.daily.dev/posts/NGytU6zYE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Global cyberespionage campaign targets critical infrastructure</strong>: A sophisticated nation-state threat group, TGR-STA-1030, has been conducting a global cyberespionage campaign since January 2024, compromising government and critical infrastructure organizations in 37 countries. The group uses phishing, N-day exploits, and a new Linux kernel rootkit called ShadowGuard to exfiltrate sensitive data, posing a significant threat to global security and economic stability. <a href="https://app.daily.dev/posts/KNu0fQxJH" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Anthropic releases Claude Opus 4.6 with agent teams and expanded context</strong>: Anthropic launched Claude Opus 4.6, featuring ‘agent teams’ for parallel task execution and an expanded context window of up to 1 million tokens, now available on Amazon Bedrock, Vercel’s AI Gateway, and Google’s Vertex AI. <a href="https://app.daily.dev/feed-by-ids?id=G7Sre82A1&amp;id=Ijw4mCG7B" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI launches enterprise AI agent platform ‘Frontier’</strong>: OpenAI introduced ‘Frontier,’ an end-to-end platform for enterprises to build, deploy, and manage AI agents, allowing them to connect to external data and applications with controlled access and permissions. <a href="https://app.daily.dev/posts/TGZyay3jK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS launches new EC2 instances</strong>: AWS released new EC2 High Memory U7i-6TB instances with Intel Xeon Sapphire Rapids processors and new C8id, M8id, and R8id instances powered by custom Intel Xeon 6 processors, offering significant performance and memory bandwidth improvements. <a href="https://app.daily.dev/posts/JPt4mjvME" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ScyllaDB integrates USearch for vector search</strong>: ScyllaDB now includes vector search capabilities using the open-source USearch library, allowing storage of vector embeddings alongside structured data and leveraging ScyllaDB’s shard-per-core architecture for optimized performance. <a href="https://app.daily.dev/posts/pKNE7iWlw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft OneDrive Copilot agents generally available</strong>: Microsoft made OneDrive agents generally available, enabling users with a Microsoft 365 Copilot license to query up to 20 documents simultaneously, with agents saved as ‘.agent’ files for sharing with collaborators. <a href="https://app.daily.dev/posts/0UIAIJz1z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel launches AI SDK for application development</strong>: Laravel released an official AI SDK in beta, providing a unified API for integrating AI capabilities into Laravel applications with support for multiple AI providers like Anthropic, Gemini, and OpenAI, and features like agent building, memory, and structured outputs. <a href="https://app.daily.dev/posts/ND9RAWZ9f" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GNU Nettle 4.0 released with post-quantum cryptography support</strong>: GNU Nettle 4.0 now supports SLH-DSA (SPHINCS+), a NIST-standardized post-quantum digital signature scheme, preparing systems for quantum-resistant security. <a href="https://app.daily.dev/posts/5YsTsMWni" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic to run Super Bowl ads mocking OpenAI’s ad strategy</strong>: Anthropic is reportedly planning Super Bowl commercials to mock OpenAI’s decision to introduce advertising in ChatGPT, positioning Claude as the ad-free alternative. Because nothing says ‘we’re serious about AI’ like spending millions on a Super Bowl ad to dunk on your competitor’s business model, rather than, you know, just building better AI. Sam Altman, naturally, called it ‘dishonest.’ <a href="https://app.daily.dev/posts/TmM2mNSSP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI bot traffic surges, threatening publishers</strong>: AI bots now account for a significant portion of web traffic, with one bot visit for every 31 human visits, primarily driven by RAG bots. Publishers are deploying ‘increasingly aggressive defenses’ against these bots, which are ‘developing sophisticated tactics to bypass website protections.’ It’s an ‘arms race,’ apparently, where AI is simultaneously the savior and destroyer of the internet, depending on who you ask. <a href="https://app.daily.dev/posts/5O2tzW5oh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Elixir is the ‘best language for AI,’ according to an unverified study</strong>: An article claims Elixir is the ‘best language for AI,’ citing an unnamed ‘Tencent study’ that supposedly found Elixir achieved a 97.5% LLM completion rate. Because a single, unverified study from an unnamed source is all you need to declare a language ‘the best’ for anything, especially when it aligns perfectly with your existing biases. Who needs peer review when you have a bold claim? <a href="https://app.daily.dev/posts/1D7tGyRwk" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
