<h2>Must know</h2>
<ul>
<li><strong>Anthropic and figma team up for code-to-design workflow</strong>: Anthropic and Figma have partnered to create a new workflow that allows Claude Code to convert production code directly into editable Figma designs. This means developers can send a live web page to a Figma file with a simple prompt, facilitating design collaboration. The updated designs can then be round-tripped back to code, streamlining the iteration process, as reported by Anthropic. <a href="https://app.daily.dev/posts/aCaE5xJ9z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent security is a growing concern</strong>: AI coding agents, while boosting productivity, are creating a ‘security debt crisis’ by prioritizing speed over secure practices, leading to issues like hardcoded API keys and public access policies. Research from Columbia University highlights this, and a security researcher demonstrated how ChatGPT recommended a malicious site, leading to a reverse C2 shell installation. Mitigations include spec-driven prompting with OWASP Top 10 and thorough code reviews. <a href="https://app.daily.dev/feed-by-ids?id=DD6LZnw0Q&amp;id=Et2d8uHu3&amp;id=LUO8dPKde&amp;id=RH1TYGIET&amp;id=UGo4Ea1t8&amp;id=bDJrJKMJX&amp;id=em7afQ5sm&amp;id=l0KLAjYGc&amp;id=ohhVPexbM&amp;id=vXBWwcwBk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s gemini 3.1 pro: high benchmarks, low usability</strong>: Despite achieving top benchmark scores, Google’s Gemini 3.1 Pro consistently underperformed Claude Opus 4.6 in real-world coding and UI generation tasks. Issues include tool-calling failures, incomplete features, and poor environment detection, leading to significant practical usability problems for developers. This suggests Google might be ‘benchmaxing’ at the expense of real-world application, according to multiple reports. <a href="https://app.daily.dev/feed-by-ids?id=bIAFUawfa&amp;id=mGa3eARS3" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>WebMCP enables browser-native AI agent execution</strong>: Google’s WebMCP is a new browser-native API allowing websites to expose structured JavaScript tools for AI agents to invoke directly within the browser, reducing token overhead and improving determinism. <a href="https://app.daily.dev/posts/61GmU0Tcy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent orchestration and memory systems evolve</strong>: Mission Control is an open-source AI agent orchestration dashboard, while LangSmith Agent Builder uses a filesystem-based memory system for persistent, task-specific agent knowledge. <a href="https://app.daily.dev/feed-by-ids?id=LY7p8IODB&amp;id=ZmcjIu8ai&amp;id=frjHTf2Ap&amp;id=xhDI1zj4i&amp;id=9xLrRFK2b" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NTransformer optimizes LLM inference on consumer gpus</strong>: NTransformer, an open-source C++/CUDA engine, enables running Llama 3.1 70B on a single RTX 3090 by streaming layers and using speculative decoding, claiming an 83x speedup over mmap baselines. <a href="https://app.daily.dev/feed-by-ids?id=Ahx4zgQks&amp;id=WoJMuyqAQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong><a href="http://Lobste.rs" target="_blank" rel="noopener nofollow">Lobste.rs</a> attempts sqlite migration, then reverts</strong>: <a href="http://Lobste.rs" target="_blank" rel="noopener nofollow">Lobste.rs</a> briefly migrated from MariaDB to SQLite but quickly reverted due to severe performance degradation, with plans to retry after addressing the issues. <a href="https://app.daily.dev/posts/RLQeRSeS0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Expert elixir language server reaches release candidate</strong>: Expert, the new official Elixir language server merging Lexical and Next LS, has released its first release candidate, focusing on stability and feature parity. <a href="https://app.daily.dev/posts/ItEtgwo7d" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Database design starts with consequences, not tables</strong>: Good database design prioritizes understanding business domain consequences, emphasizing rigorous schema changes, deliberate denormalization, and using constraints to encode business rules. <a href="https://app.daily.dev/posts/fI73pTtiz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Android malware uses advanced persistence techniques</strong>: A deep dive into Android malware reveals sophisticated persistence mechanisms, including boot persistence, crash recovery, remote resurrection via FCM, and Device Administrator privileges to prevent uninstallation. <a href="https://app.daily.dev/posts/VKUMk8xIJ" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Meta patents ai to post after you’re dead</strong>: Meta apparently thinks we all want our social media presence to continue after we’re gone, patenting an AI system that simulates deceased users’ activity. Because nothing says ‘grieving process’ like your dead aunt’s AI posting about her favorite cat memes. Matt Wolfe reports on this groundbreaking innovation in digital necromancy. <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenClaw: the ai agent for everything, except permissions</strong>: OpenClaw, an autonomous AI agent, promises to handle everything from automated dependency maintenance to cold email lead generation. But wait, it can’t respond to Claude Code permission prompts autonomously? So, it’s ‘autonomous’ until it hits a basic security check, then it needs a human or you just ‘dangerously-skip-permissions.’ Sounds like a solid plan for enterprise adoption. <a href="https://app.daily.dev/posts/WCj2dFu7Q" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Adobe and nvidia: 10,000,000 sparkles at 280 fps</strong>: Adobe and NVIDIA have teamed up to render millions of microscopic glint particles at 280 FPS. Because what every user-centric experience truly needs is more sparkles. Forget robust system design or efficient development; we’re optimizing for glitter, apparently. Two Minute Papers highlights this critical advancement in visual effects. <a href="https://app.daily.dev/posts/rT496vKCA" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
