<h2>Must know</h2>
<ul>
<li><strong>NVIDIA pushes AI gaming and enterprise with new hardware and software</strong>: NVIDIA announced DLSS 4.5 with Dynamic Multi Frame Generation for 240+ FPS path-traced gaming on RTX 50 Series GPUs, and G-SYNC Pulsar monitors for over 1,000Hz motion clarity. For enterprise, NVIDIA expanded its AI Factory validated design with BlueField cybersecurity and infrastructure acceleration, integrating with nine ecosystem partners. They also unveiled DGX Spark and DGX Station, deskside AI supercomputers for local large-scale AI model development, as reported by NVIDIA. <a href="https://app.daily.dev/feed-by-ids?id=jXjzYY9vI&amp;id=bXoqeH8eP&amp;id=SCEwsKfyS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Major vendors ditch MITRE ATT&amp;CK evaluations</strong>: Microsoft, SentinelOne, and Palo Alto Networks have pulled out of the 2025 MITRE ATT&amp;CK Evaluations for endpoint detection and response, citing resource prioritization for product development. MITRE’s CTO admitted the 2025 test became more demanding, expanding to cloud environments and alert volume tracking. Industry experts are calling the process ‘vendor theater,’ suggesting it’s more for PR than genuine security improvement. <a href="https://app.daily.dev/feed-by-ids?id=c0ZYfkcgP&amp;id=hu11lcW1J" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Mandiant releases NTLMv1 rainbow tables to force deprecation</strong>: Mandiant publicly released rainbow tables capable of cracking Net-NTLMv1 password hashes in under 12 hours using consumer hardware. This move aims to accelerate the deprecation of the 20-year-old, vulnerable protocol, which Mandiant says is still widely used despite Microsoft’s recommendations. They emphasize that credential theft is trivial with NTLMv1 and urge organizations to disable it immediately, according to their official release. <a href="https://app.daily.dev/posts/2q2phjvWS" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>OpenAI’s GPT-5.2-Codex now available in Responses API</strong>: OpenAI has released GPT-5.2-Codex in its Responses API, enhancing capabilities for complex software engineering tasks like feature building, code refactoring, and bug finding, with improved long-context understanding and tool-calling. <a href="https://app.daily.dev/posts/lZuQnCdnA" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pulumi adds native support for Terraform and HCL</strong>: Pulumi now natively supports Terraform and HCL, allowing direct HCL execution and Terraform state management within Pulumi Cloud, with a private beta currently available and general availability expected in Q1 2026. <a href="https://app.daily.dev/posts/brDXTtnxJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare automates Salt configuration debugging</strong>: Cloudflare developed Jetflow, an event-driven observability framework, to automate debugging of SaltStack configuration errors across millions of servers, reducing release delays by 5% and eliminating hours of manual log analysis. <a href="https://app.daily.dev/posts/s63JTpz4h" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rails updates include error-aware retries and PostgreSQL types</strong>: Rails now features error-aware retry strategies in ActiveJob, improved CSRF protection, removal of the deprecated X-XSS-Protection header, and a new public API for PostgreSQL custom type mappings. <a href="https://app.daily.dev/posts/LkAtCVwbE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Chrome 145 beta brings CSS and Web API enhancements</strong>: Chrome 145 beta introduces new CSS features like text-justify and column wrapping, along with Web API updates including Navigation API destination exposure, WebGPU subgroup uniformity, and ECMAScript Map upsert methods. <a href="https://app.daily.dev/feed-by-ids?id=oWgZOWIXh&amp;id=opNGD5YZ9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Shopify deprecates tracking cookies for Hydrogen storefronts</strong>: Shopify is deprecating <code>shopify_y</code> and <code>shopify_s</code> tracking cookies in 2026, requiring Hydrogen storefronts to upgrade by April 30, 2026, to maintain reliable visitor and session attribution in Shopify Analytics. <a href="https://app.daily.dev/posts/QzarOUOVo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New interactive load balancer simulator released</strong>: An interactive simulator tool has been released, allowing users to visualize load balancer traffic distribution across multiple servers and test different algorithms like Round Robin, Random, and Least Connections in real-time. <a href="https://app.daily.dev/feed-by-ids?id=SEEQH14TR&amp;id=wXih2mfvP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-generated code: the gift that keeps on giving… technical debt</strong>: Multiple engineering teams are reporting that AI-generated code, while compiling initially, often lacks architectural context and creates poor state management, leading to major refactoring that can take longer than writing it from scratch. Apparently, trusting ‘plausible-looking’ AI outputs without expert oversight is a fantastic way to embed silent technical debt, as ChatGPT’s ‘realistic’ but non-existent API elements have shown. Who knew that a tool generating code without understanding the bigger picture might cause problems? <a href="https://app.daily.dev/feed-by-ids?id=STWE5tRXe&amp;id=UAphs9F5B&amp;id=euzGVQGzQ&amp;id=gMHZMigM0&amp;id=m6P4U0J0b&amp;id=wCQfiCXDa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents: just APIs with a fancy new hat</strong>: According to some, ‘AI agents’ are just a rebranding of API-centric automation, primarily driven by new rounds of investment rather than genuine technical breakthroughs. While developers are enthusiastically building Pokémon Trainer AIs with LangChain and using Claude Code for code migration, the core argument is that the foundational concepts have been around for over a decade. So, next time someone talks about ‘agentic technology,’ remember it might just be your old friend, the API, wearing a new, expensive suit. <a href="https://app.daily.dev/feed-by-ids?id=4rJ96gwiU&amp;id=6IAZSWIz1&amp;id=6XeVWOBXO&amp;id=77jDjhI5C&amp;id=sH4hATTfN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Claude Co-work: AGI’s buggy desktop debut</strong>: Anthropic’s Claude Co-work, a desktop automation tool for non-developers, is being touted as a ‘small taste of AGI.’ However, early reports suggest it’s more like a small taste of frustration, with broken authentication, UI glitches, and confusing permission models. While it can organize files and execute terminal commands, it struggles with cloud storage and has security concerns around prompt injection. So, while the underlying tech is powerful, the user experience seems to be stuck in beta hell, proving that even AGI needs a decent UX team. <a href="https://app.daily.dev/posts/OemgIV2rY" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
