<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 for enterprises using AI agents, affecting platforms like OpenClaw and Claude Code. Operant AI launched Agent Protector to offer real-time security for autonomous AI agents, focusing on shadow agent discovery and inline protection against privilege escalation and data exfiltration. <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>Substack user data breach disclosed</strong>: Substack disclosed a data breach from October 2025 that exposed user email addresses, phone numbers, and internal metadata, remaining undetected for four to five months. While passwords and financial information were not compromised, a threat actor claims nearly 700,000 records were affected. Substack has fixed the vulnerability and is investigating, but the delay in detection and lack of transparency raise significant privacy concerns. <a href="https://app.daily.dev/posts/myrJIoXYs" 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 in 37 countries. The group uses phishing, N-day exploits, and a new Linux kernel rootkit called ShadowGuard to exfiltrate sensitive data. This ongoing threat underscores the need for robust cybersecurity measures and vigilance against advanced persistent threats. <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</strong>: Anthropic launched Claude Opus 4.6, featuring ‘agent teams’ for parallel task execution, a 1 million token context window, and new integrations with Amazon Bedrock, Vercel, and Google 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. <a href="https://app.daily.dev/posts/TGZyay3jK" 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 and Windows patch problems. <a href="https://app.daily.dev/posts/NGytU6zYE" 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 from multiple providers like Anthropic, Gemini, and OpenAI into Laravel applications. <a href="https://app.daily.dev/posts/ND9RAWZ9f" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS launches new EC2 instances</strong>: AWS launched new EC2 High Memory U7i-6TB instances and C8id, M8id, and R8id instances powered by custom Intel Xeon 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 for optimized performance in AI applications. <a href="https://app.daily.dev/posts/pKNE7iWlw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Amazon ECS adds NLB deployment support</strong>: Amazon ECS now supports linear and canary deployment strategies for services using Network Load Balancers, enabling controlled, incremental traffic shifting for TCP/UDP-based applications. <a href="https://app.daily.dev/posts/4MU5wceHx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s super bowl ad stunt</strong>: Anthropic is planning Super Bowl ads to mock OpenAI for introducing ads in ChatGPT, positioning Claude as the ‘ad-free’ AI. OpenAI’s Sam Altman, naturally, called it ‘dishonest.’ It’s a classic tech rivalry, but let’s be real, both companies are just trying to figure out how to make money off these expensive toys. Anthropic’s ‘ad-free’ stance feels a lot like a temporary marketing gimmick until they need to burn through more billions. <a href="https://app.daily.dev/posts/TmM2mNSSP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI bot traffic surges, publishers panic</strong>: Apparently, AI bots now account for one web visit for every 31 human visits, and publishers are freaking out because clickthrough rates from AI apps have plummeted to 0.27%. The internet is ‘reshaping’ because AI users are going directly to AI platforms instead of traditional search. So, the internet is becoming a giant, unmonetized data source for AI, and publishers are left wondering why their ad revenue is tanking. Who could have seen that coming? <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’ says unverified study</strong>: An article boldly 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 nothing screams credibility like an anonymous study from a major tech company that can’t be found anywhere else. It’s almost as if someone is trying to hype up their favorite language with some conveniently unverified data. <a href="https://app.daily.dev/posts/1D7tGyRwk" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
