<h2>Must know</h2>
<ul>
<li><strong>Google Gemini AI faces extraction attempts and malicious use</strong>: Google reported blocking over 100,000 attempts to extract proprietary reasoning capabilities from its Gemini AI model, which they consider intellectual property theft. Nation-state actors from China, Iran, North Korea, and Russia are also integrating Gemini into attack operations for social engineering and phishing, with new malware families like HONESTCUE embedding Gemini’s API to generate and execute malicious code, bypassing safety filters. This highlights the dual-use nature of advanced AI and the escalating cybersecurity risks. <a href="https://app.daily.dev/posts/OQLlwLpSw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.0 kernel updates and post-quantum cryptography integration</strong>: The upcoming Linux 7.0 kernel is bringing significant performance optimizations, including improved memory management and XFS file system enhancements like autonomous self-healing. Crucially, it integrates ML-DSA, a quantum-resistant signature algorithm, for module signing, replacing SHA-1 support. This move towards post-quantum cryptography, standardized by NIST, is a proactive step against future quantum computing threats to current cryptographic algorithms like RSA and ECDSA. <a href="https://app.daily.dev/feed-by-ids?id=6DJe6t3uU&amp;id=Ukr75F53X&amp;id=Vay0ocYOF&amp;id=Vh5TboDqs&amp;id=m5L4rWAXQ&amp;id=ppwBwOuN5&amp;id=uGdXnDZH4&amp;id=xVcPYGmKq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>TypeScript 6.0 beta released, Go compiler transition confirmed</strong>: TypeScript 6.0 beta is now available, marking the final version built on its JavaScript codebase. The big news is that TypeScript 7.0 will transition to a Go-based compiler, aiming for improved performance and efficiency. This beta includes a <code>--stableTypeOrdering</code> flag to help with migration, support for <code>es2025</code> and new Temporal API types, and improvements to <code>this-less</code> function handling. <a href="https://app.daily.dev/feed-by-ids?id=62kNn8zlZ&amp;id=vVvcWvIPI" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI’s impact on software development and developer roles</strong>: AI tools are eliminating routine coding tasks, shifting the bottleneck from writing code to architectural decisions and verification, making junior developer roles more challenging while boosting senior engineers’ productivity. <a href="https://app.daily.dev/feed-by-ids?id=0wFgrbFqP&amp;id=78oOrNVDz&amp;id=B3N2Icy7Y&amp;id=BKJ8NHw6d&amp;id=DcSKM9BQA&amp;id=GManH4sq3&amp;id=GnwSkM7qk&amp;id=HHMxTl4yN&amp;id=JgHE2Ywe8&amp;id=MPMZLgvp3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Go 1.26 released with performance improvements and new static analyzer</strong>: Go 1.26 is out, featuring the Green Tea garbage collector by default, a 30% reduction in <code>cgo</code> call overhead, experimental SIMD support, and a new static analyzer from PVS-Studio that claims to catch bugs <code>go vet</code> misses. <a href="https://app.daily.dev/feed-by-ids?id=1ObNpwyIZ&amp;id=mj2oUwkea" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New AI models and hardware shifts for faster inference</strong>: OpenAI released GPT-5.3-Codex-Spark, a coding model running on Cerebras CS3 accelerators for 15x faster inference, while <a href="http://Z.ai" target="_blank" rel="noopener nofollow">Z.ai</a> launched GLM5, another fast coding model, and ByteDance introduced Seedance 2.0 for multimodal video generation. <a href="https://app.daily.dev/feed-by-ids?id=2DKu3o06c&amp;id=2G5oHg0oJ&amp;id=A9s0KznyU&amp;id=CacFMCP2L&amp;id=PFJ429xnc&amp;id=TgVTzT7ZG&amp;id=dqsxZt8l5&amp;id=f8bwmVFOm&amp;id=iVwXrkdSj&amp;id=yYXkpJGbb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>PostgreSQL extension for ClickHouse metrics</strong>: ClickHouse open-sourced <code>pg_stat_ch</code>, a PostgreSQL extension that streams every query execution as a ~4.6KB event to ClickHouse for APM-style analysis with minimal overhead, capturing 45 metrics per query. <a href="https://app.daily.dev/posts/UtmEKBs89" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Self-hosting ecosystem developments</strong>: Orchestra, a new open-source PaaS, automates bare metal server provisioning and application deployment, while GoClaw offers an ultra-lightweight AI assistant for $10 hardware, and Caddy web server continues to impress with automatic SSL and performance. <a href="https://app.daily.dev/feed-by-ids?id=0nUdXf1b9&amp;id=C2t5MyKnp&amp;id=jNBhBoH4V&amp;id=nwGZvk6OS&amp;id=sjdf7M4Ia&amp;id=qxYADTM4h" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>HBM4 memory enters mass production for AI accelerators</strong>: Samsung and Micron have begun mass production and shipping of HBM4 memory, the next-gen high-bandwidth memory for AI accelerators, with Samsung’s HBM4 delivering 11.7 Gbps and 40% better energy efficiency than HBM3E. <a href="https://app.daily.dev/feed-by-ids?id=OUBYQ2PHL&amp;id=WTPyAW1Ca&amp;id=gpkg8awYO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust leadership council elections and open-source updates</strong>: The Rust Leadership Council is initiating its March 2026 representative selection process, while openSUSE Board member interviews are underway, and KDE’s immutable Linux distribution is nearing completion. <a href="https://app.daily.dev/feed-by-ids?id=7UGQouctP&amp;id=Hs3HWV6zV" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI will fix all your cloud operations, just trust Microsoft</strong>: Microsoft is pushing ‘agentic cloud operations’ where AI-powered agents supposedly optimize your cloud environments continuously. Azure Copilot is their unified interface for deploying these specialized agents across the entire cloud lifecycle. Because, you know, what could possibly go wrong with autonomous AI agents making changes to your production infrastructure? What could possibly go wrong? <a href="https://app.daily.dev/feed-by-ids?id=IQCKAyNnH&amp;id=XwFeQESAO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is making security more agile, or just more complicated</strong>: ChiBrrCon 2026 highlighted how AI is transforming enterprise security, emphasizing ‘operational agility’ over traditional resilience. Speakers stressed building adaptability through stress-tested runbooks and context-rich alerts. So, AI is supposed to make us more agile, but we still need to stress-test everything and have runbooks for when the AI inevitably goes sideways. Sounds like more work, not less. <a href="https://app.daily.dev/posts/FyP2Wb8Vm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are the future, just ignore the data quality problem</strong>: Elastic is telling us that AI agents need ‘accurate, well-organized, context-rich data’ to function effectively in enterprises, but then admits enterprises struggle with ‘messy, unstructured information.’ Their solution is an ‘Agent Builder framework’ to address these data challenges. So, the problem is bad data, and the solution is… another tool to manage the bad data for the AI that needs good data. Got it. <a href="https://app.daily.dev/feed-by-ids?id=IQCKAyNnH&amp;id=XwFeQESAO" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
