<h2>Must know</h2>
<ul>
<li><strong>Stripe considers acquiring PayPal</strong>: Stripe is reportedly eyeing PayPal for a potential acquisition, or at least parts of it, according to Bloomberg. This comes as Stripe’s valuation soared to $159 billion, a 74% jump from last year, while PayPal, valued at over $40 billion, has been struggling with slowing growth. The talks are early, but the market reacted positively to PayPal’s stock, suggesting investors see potential in such a merger. <a href="https://app.daily.dev/feed-by-ids?id=gB7UDBJXo&amp;id=2WYO6dKLA&amp;id=HqWjbO0Bb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cybersecurity landscape worsens with AI-enabled attacks</strong>: CrowdStrike’s latest report paints a grim picture: the average time for a cyberattack to ‘break out’ dropped to just 29 minutes, a 65% decrease year-over-year. AI-enabled attacks surged by 89%, with adversaries using generative AI across the entire attack chain. This shift means more stealthy, malware-free intrusions, highlighting a critical need for proactive threat hunting and comprehensive AI governance. <a href="https://app.daily.dev/posts/WsqeDq8OW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>npm supply chain attack targets AI coding tools</strong>: Socket researchers uncovered a new npm supply chain attack, ‘SANDWORM_MODE,’ involving 19 malicious packages typosquatting popular developer and AI coding tools, including Claude Code. These packages steal credentials like npm/GitHub tokens and cloud secrets, then inject malicious code into repositories and CI pipelines. They even deploy a malicious MCP server for prompt-injection attacks against AI coding assistants, silently exfiltrating SSH keys and cloud credentials. <a href="https://app.daily.dev/posts/q6c2RLQKH" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Apple researchers unveil on-device AI model for UIs</strong>: Apple researchers introduced Ferret-UI Lite, a compact 3B-parameter multimodal AI model designed to understand and interact with user interfaces directly on-device, prioritizing low latency, privacy, and offline capability. <a href="https://app.daily.dev/posts/qj9GoBvQx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New LLM inference technique triples speed</strong>: Researchers developed a multi-token prediction technique that triples LLM inference speed without auxiliary draft models, achieving 3x+ speedup on math benchmarks with under 3% accuracy loss by fine-tuning models with a confidence-adaptive decoding strategy. <a href="https://app.daily.dev/posts/hxaAjvILy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents accelerate feature development on personal site</strong>: Kent C. Dodds reported using AI agents, specifically Cursor, to rapidly implement features like semantic search and keyboard shortcuts on his personal site, clearing a backlog of GitHub issues. <a href="https://app.daily.dev/posts/HVH19Oob1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cursor launches autonomous cloud AI agents for software development</strong>: Cursor released cloud agents that autonomously develop software in isolated virtual machines, producing merge-ready Pull Requests with video and screenshot artifacts, and allowing parallel execution of long-running tasks. <a href="https://app.daily.dev/posts/BbESdM4Ct" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quesma releases OTelBench for OpenTelemetry and AI performance</strong>: Quesma launched OTelBench, an open-source benchmarking suite to evaluate OpenTelemetry pipeline performance and AI agent effectiveness in SRE tasks, revealing that even frontier LLMs often achieve success rates below 30% in production-grade instrumentation. <a href="https://app.daily.dev/posts/OB1MlP8PG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Red Hat launches AI enterprise platform</strong>: Red Hat introduced Red Hat AI Enterprise, a platform designed to move AI workloads from pilot to production by providing a standardized environment for inference, model customization, and agent deployment, supporting any model or hardware. <a href="https://app.daily.dev/posts/e1iAecIls" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent infrastructure, security, and observability developments</strong>: New developments in AI agent infrastructure include Agentuity’s Gravity Network for ephemeral agent workloads, Tinybird’s GA Branches for isolated data environments, and New Relic’s Agentic Platform for data observability, all addressing security and reliability for AI agents. <a href="https://app.daily.dev/feed-by-ids?id=0HG7ge0iD&amp;id=29JQKm5zh&amp;id=4Ak1fvjod&amp;id=EUE0Dit8s&amp;id=F9Y7SyAcI&amp;id=OBZ9FuH7A&amp;id=fWsa69inU" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AWS blames engineer for AI tool outage, not the AI tool</strong>: AWS published a blog post blaming an engineer with ‘broader permissions than expected’ for accidentally triggering a CloudFormation teardown in production using their Kiro AI coding tool. AWS called Kiro’s involvement a ‘coincidence,’ which is a convenient way to protect their AI’s reputation while simultaneously laying off thousands of human workers. The fix? Mandatory human peer review for AI-generated changes. So much for AI autonomy. <a href="https://app.daily.dev/posts/qZoHoApRh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents make the ‘10x developer’ debate moot</strong>: Justin Searls, a named individual, claims AI agents have rendered the ‘10x developer’ debate moot, effectively equalizing or amplifying developer output across the board. This is a bold, self-serving claim with zero evidence provided on how AI agents achieve this magical equalization. Sounds like someone’s trying to justify their AI investment, or perhaps their own job. <a href="https://app.daily.dev/posts/dMpIU74OH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Samsung’s ‘upcycling’ program: greenwashing or genuine effort?</strong>: Samsung won a sustainability award for its ‘Galaxy Upcycling’ program, which initially promised bootloader unlocking and custom OS support for old phones. Four years later, the program launched with a drastically reduced scope, offering only basic sound detection and brightness measurement. iFixit called it ‘nearly unrecognizable.’ Getting an award for a program that barely delivered on its promises? That’s some impressive greenwashing. <a href="https://app.daily.dev/posts/ebOB9DlFn" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
