<h2>Must know</h2>
<ul>
<li><strong>AI agent security vulnerabilities are critical</strong>: Snyk’s research, ‘ToxicSkills,’ found that 13.4% of AI agent skills have critical security issues, with 76 confirmed malicious payloads designed for credential theft and data exfiltration. This highlights a significant supply chain risk in the AI agent ecosystem, affecting platforms like OpenClaw and Claude Code. Operant AI launched Agent Protector to provide real-time security for autonomous AI agents, focusing on preventing privilege escalation and data exfiltration, emphasizing that security should be embedded by platform providers, not left to 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>Anthropic and OpenAI escalate AI chatbot ad war</strong>: Anthropic plans Super Bowl ads mocking OpenAI’s move to introduce advertising in ChatGPT, positioning its Claude AI as ad-free. OpenAI CEO Sam Altman called Anthropic’s ads ‘dishonest,’ defending OpenAI’s ad principles. This public spat highlights differing business models and intense rivalry, with Anthropic relying on enterprise contracts while OpenAI reportedly burns $9 billion annually with only 5% paying users. <a href="https://app.daily.dev/posts/TmM2mNSSP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Global cyberespionage campaign targets critical infrastructure</strong>: A sophisticated nation-state group, TGR-STA-1030, has been conducting a global cyberespionage campaign since January 2024, compromising government and critical infrastructure in 37 countries. The group uses phishing, N-day exploits, and a new Linux kernel rootkit, ShadowGuard, to exfiltrate sensitive data. This widespread campaign underscores the escalating threat landscape and the need for robust cybersecurity measures across sectors. <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 Opus 4.6, featuring ‘agent teams’ for parallel task execution, a 1 million token context window, and new PowerPoint integrations, aiming to broaden its appeal beyond developers to knowledge workers. <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, treating them like employees with onboarding and feedback loops, and aiming to connect siloed data systems. <a href="https://app.daily.dev/posts/TGZyay3jK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding agents advance software development automation</strong>: AI coding agents are automating tasks from feature requests to deployment, with Claude Code introducing agent teams, Cursor achieving 1,000 commits per hour, and LLMs rewriting parsers, though requiring significant human refinement. <a href="https://app.daily.dev/feed-by-ids?id=I2lSMdhWD&amp;id=Q3sR9E9tC&amp;id=WcVKYHvlb&amp;id=bgc1cUCVA&amp;id=bq4ihvY3K&amp;id=dsKDEbeOE&amp;id=RtOeT8YDy&amp;id=YPOCeTWKC&amp;id=gdolSqfBx&amp;id=kPp1hKs76" 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 engineering quality czar, reporting directly to him, to address ongoing issues like Azure outages, Windows patch problems, and low Copilot adoption rates. <a href="https://app.daily.dev/posts/NGytU6zYE" 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>Rust ecosystem update #637</strong>: The latest ‘This Week in Rust’ newsletter highlights new library additions, stabilized features like <code>if let</code> guards, improved compiler performance, and community articles on network programming and SIMD benchmarking. <a href="https://app.daily.dev/posts/HbfZ0FSX1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Elixir is the ‘best language for AI’</strong>: An article claims Elixir is the ‘best language for AI,’ citing an unverified ‘Tencent study’ that supposedly found Elixir achieved a 97.5% LLM completion rate. Apparently, Elixir’s immutability and ‘decade-long stability’ are the secret sauce for AI, because who needs actual benchmarks or peer-reviewed studies when you have a ‘Tencent study’ and a strong pro-Elixir bias? <a href="https://app.daily.dev/posts/1D7tGyRwk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI bot traffic surges, threatening publishers</strong>: AI bots now constitute a significant portion of web traffic, with one bot visit for every 31 human visits, and clickthrough rates from AI applications plummeting to 0.27%. Publishers are deploying ‘increasingly aggressive defenses’ against these bots, which are ‘developing sophisticated tactics to bypass website protections.’ It’s an ‘arms race,’ folks, where the internet is being reshaped by AI that can’t even bother to click a link. The future is here, and it’s a bot scraping your content without attribution. <a href="https://app.daily.dev/posts/5O2tzW5oh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>UK deepfake detection plan criticized</strong>: The UK government announced plans to develop a framework for evaluating deepfake detection technologies, partnering with Microsoft and academics. Cybersecurity expert Dr. Ilia Kolochenko, however, states this initiative is ‘unlikely to produce systemic improvements,’ arguing that existing open-source tools already track AI-generated content. Because, clearly, the problem isn’t detecting deepfakes; it’s the ‘enforcement’ part, which a government framework will surely solve. <a href="https://app.daily.dev/posts/EQCYBDiDB" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
