<h2>Must know</h2>
<ul>
<li><strong>Stripe considers acquiring PayPal</strong>: Stripe is reportedly eyeing PayPal for a potential acquisition, according to Bloomberg. This comes as Stripe’s valuation recently surged to $159 billion, a 74% increase from last year, while PayPal, valued at over $40 billion, has been struggling with slowing growth. The talks are very early, but the news sent PayPal’s shares up nearly 7%. <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>Major npm supply chain attack targets AI coding tools</strong>: Socket researchers uncovered the ‘SANDWORM_MODE’ npm supply chain attack, involving 19 typosquatted packages impersonating developer utilities and AI coding tools like Claude Code. The malware harvests credentials, injects malicious code into CI pipelines via GitHub Actions, and deploys a malicious MCP server for prompt-injection attacks, silently exfiltrating SSH keys and cloud credentials. This highlights critical vulnerabilities in AI-driven development workflows. <a href="https://app.daily.dev/feed-by-ids?id=q6c2RLQKH&amp;id=3fX3MdYNS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cybersecurity landscape worsens with AI-enabled attacks</strong>: CrowdStrike’s 2026 Global Threat Report indicates a dramatically worsening cybersecurity landscape, with average cyberattack breakout time dropping to 29 minutes. AI-enabled attacks surged 89%, with adversaries integrating Generative AI across the attack kill chain. The report notes a shift towards stealthy, malware-free intrusions and a 37% rise in cloud intrusions, emphasizing the 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>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Meta secures major AMD chip deal for AI expansion</strong>: Meta signed a multiyear deal with AMD for up to $100 billion in chips, including MI540 GPUs, to add 6GW of data center compute capacity and diversify its AI compute stack beyond Nvidia. <a href="https://app.daily.dev/posts/vYZNsZv8f" 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 with under 3% accuracy loss on reasoning-heavy tasks 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 for autonomous software development launched</strong>: Cursor launched cloud agents for autonomous software development, operating in isolated VMs with full dev environments, capable of producing merge-ready Pull Requests with video and screenshot artifacts, and running multiple long-running tasks in parallel. <a href="https://app.daily.dev/posts/BbESdM4Ct" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apple researchers introduce on-device AI model Ferret-UI Lite</strong>: Apple researchers unveiled Ferret-UI Lite, a compact 3B-parameter multimodal AI model designed for on-device UI understanding and interaction, prioritizing low latency, privacy, and offline capability, potentially reducing Siri’s cloud dependence. <a href="https://app.daily.dev/posts/qj9GoBvQx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hetzner Cloud price hike due to hardware costs</strong>: Hetzner, a major European budget cloud provider, announced price increases of up to 50% starting April 1, 2026, citing dramatic rises in hardware costs driven by AI industry demand, which disproportionately affects mid-tier providers without long-term vendor contracts. <a href="https://app.daily.dev/posts/van1jyyDC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-augmented infrastructure audits reduce prep time by 80%</strong>: A case study showed AI agents, specifically Claude Code with MCP servers, reduced audit preparation time by 80% for an AWS CDK infrastructure stack, auto-generating a full CloudWatch monitoring layer, though 30% of recommendations were contextually incorrect. <a href="https://app.daily.dev/posts/SgcSx3o2T" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Enveil protects .env secrets from AI coding tools</strong>: GreatScott released <code>enveil</code>, a Rust CLI tool that replaces plaintext secrets in <code>.env</code> files with symbolic references, storing real values in an encrypted local store and injecting them directly into subprocess environments at runtime, protecting them from AI coding tools. <a href="https://app.daily.dev/posts/sSPQXaHKu" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AWS blames engineer for AI tool outage</strong>: AWS published a blog post blaming an engineer with ‘broader permissions than expected’ for accidentally triggering a CloudFormation teardown in production using their AI coding tool, Kiro. AWS framed Kiro’s involvement as a ‘coincidence,’ which is a convenient way to protect their AI’s reputation while simultaneously laying off thousands of human workers. The solution? Mandatory peer review of AI-generated changes by human engineers. So, AI makes mistakes, humans clean them up, and AWS still cuts human jobs. Peak efficiency, right? <a href="https://app.daily.dev/posts/qZoHoApRh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents make ‘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 bold assertion, framed as his prediction being ‘nailed,’ conveniently lacks any specific details or evidence on how AI agents achieve this magical equalization. It sounds more like a self-serving pat on the back than a rigorous analysis of developer productivity. <a href="https://app.daily.dev/posts/dMpIU74OH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Research and Salesforce confirm industry suspicions about AI conversations</strong>: Microsoft Research and Salesforce analyzed over 200,000 AI conversations and reportedly ‘confirmed findings the industry already suspected.’ What findings? What suspicions? They don’t say. This vague announcement, originating from a retweet, is a masterclass in saying absolutely nothing while sounding profound. It’s like announcing you’ve discovered water is wet and expecting a Nobel Prize. <a href="https://app.daily.dev/posts/FAzFsVs8B" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
