<h2>Must know</h2>
<ul>
<li><strong>AI is driving component shortages and hardware delays</strong>: The insatiable demand from AI datacenters is causing significant memory and storage component shortages, impacting the broader hardware industry. Valve, for instance, is experiencing intermittent stock issues for Steam Deck OLED units and has delayed future hardware launches, including the Steam Machine, due to these shortages. This trend could even push back the release of consoles like the PlayStation 6 until 2028-2029, as reported by Valve. <a href="https://app.daily.dev/posts/ttFo5ZiRW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>European Parliament bans cloud AI on devices over data security</strong>: The European Parliament has disabled cloud-based AI features on lawmakers’ corporate devices due to serious data security and privacy concerns. Their IT department cannot guarantee data security when AI assistants use cloud services, fearing confidential information could be processed in unknown cloud locations, especially given many AI company servers are U.S.-based. This temporary ban highlights institutional unease about AI handling sensitive legislative data, despite the EU’s own comprehensive AI Act. <a href="https://app.daily.dev/posts/NW6nIRqB9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are being exploited as covert C2 proxies</strong>: Check Point Research demonstrated how AI assistants like Grok and Microsoft Copilot can be exploited as covert Command-and-Control (C2) proxies without API keys or accounts, allowing attackers to tunnel malware communications. This means AI models can act as runtime decision engines for victim triage, data prioritization, and sandbox evasion, accelerating reconnaissance and data exfiltration. This blends malicious traffic with legitimate enterprise AI usage, making detection harder. <a href="https://app.daily.dev/feed-by-ids?id=8kxTLeIWY&amp;id=DCxMGa8ig&amp;id=H92K14NPx&amp;id=JY689C6zi&amp;id=K2suxKLLr&amp;id=T8PHWMs0E" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>MongoDB Java driver sees significant performance gains</strong>: MongoDB’s Java team achieved 20-90% throughput improvements in their Java driver by optimizing hot paths, including caching BSON array indexes and leveraging JVM intrinsics for byte order, emphasizing measurement over intuition. <a href="https://app.daily.dev/feed-by-ids?id=0JRJsAJbi&amp;id=6dgAlz6XB&amp;id=C9og01r53&amp;id=Ii1csxjgn&amp;id=vkFxQXGK8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.0 optimizes close_range system call</strong>: Linux 7.0 introduces a performance optimization for the <code>close_range</code> system call, reducing its complexity from O(range size) to O(active file descriptors) by using <code>find_next_bit()</code> on the <code>open_fds</code> bitmap, significantly speeding up operations on sparse file descriptor tables. <a href="https://app.daily.dev/posts/2IrhD5aAX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS SageMaker supports custom Amazon Nova models</strong>: AWS now supports deploying custom Amazon Nova models (Micro, Lite, Nova 2 Lite) through Amazon SageMaker Inference, enabling production-grade deployments with configurable instance types, auto-scaling, and cost optimization features. <a href="https://app.daily.dev/posts/8Nmwedcuv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Temporal secures $300M for AI and workflow orchestration</strong>: Temporal secured $300M in Series D funding at a $5B valuation, driven by demand for Durable Execution in AI and long-running workflows, with companies like OpenAI using it for AI agent orchestration. <a href="https://app.daily.dev/feed-by-ids?id=Btt27SjiL&amp;id=H860fyfMj&amp;id=HqHJFpyjl&amp;id=ME7WSXyrj&amp;id=O7fDdRA8l&amp;id=X1zVEwts3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OWASP releases Smart Contract Top 10 2026</strong>: OWASP released its Smart Contract Top 10 2026, identifying governance failures, privilege misconfiguration, and upgrade authority concentration as dominant security risks in blockchain systems, based on hundreds of millions in losses from 2025 exploits. <a href="https://app.daily.dev/posts/U7Z7c8wKd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Python Security Response Team seeks new members</strong>: The Python Security Response Team (PSRT) is actively seeking new members with security expertise to coordinate vulnerability reports and remediations for CPython and pip, formalizing its governance via PEP 811. <a href="https://app.daily.dev/posts/k8sHlR5Xv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NetBox 4.5.2 brings significant performance improvements</strong>: NetBox 4.5.2 delivers substantial performance improvements, including 60% faster REST API response times through request-level caching and cursor-based GraphQL pagination, particularly benefiting deployments managing millions of network infrastructure objects. <a href="https://app.daily.dev/posts/RhwD6wBej" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Alibaba’s Qwen3.5 claims to beat GPT-5.2, Claude 4.5 Opus, and Gemini 3 Pro</strong>: Alibaba just dropped Qwen3.5, a new multimodal AI model, claiming it outperforms GPT-5.2, Claude 4.5 Opus, and Gemini 3 Pro on benchmarks. Sure, Alibaba, sure. We’ve heard this song and dance before. It’s always the new kid on the block with the ‘benchmark improvements,’ conveniently forgetting the real-world adoption challenges like ‘trust concerns regarding its Chinese origin’ and ‘limited global ecosystem maturity.’ Good luck with those ‘structured enterprise workflows’ when nobody trusts your data residency. <a href="https://app.daily.dev/posts/UOocLbfK4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Activist investor declares toilet maker an AI play</strong>: Palliser Capital, an activist investor, is trying to convince everyone that Japanese toilet manufacturer Toto is an ‘AI play’ because its advanced ceramics division makes electrostatic chucks for 3D NAND memory chips. Apparently, this makes a toilet company an ‘overlooked AI play’ with a ‘$3.6 billion valuation gap.’ It’s a bold strategy, Cotton, let’s see if it pays off for them. Or if it’s just another desperate attempt to pump a stock by slapping ‘AI’ on it. <a href="https://app.daily.dev/posts/08TkDSLKo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI FOMO is real, and it’s making tech leaders dumb</strong>: A new report suggests that ‘AI FOMO’ is affecting over one in nine adults, and 60-70% of tech leaders are investing in AI out of fear, leading to ‘poor strategic decisions and unmeasurable returns.’ So, basically, everyone’s panicking and throwing money at AI because they’re scared of missing out, not because it makes business sense. And 74% of CIOs regret recent AI vendor choices? Sounds about right. Maybe try understanding the tech before you buy into the hype, folks. <a href="https://app.daily.dev/feed-by-ids?id=7QpmQdJRy&amp;id=ErBrh9iYd&amp;id=HVST5H2Ad&amp;id=IVuZxLuLW&amp;id=WL5gkP0JC&amp;id=hUyOtzJRD" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
