<h2>Must know</h2>
<ul>
<li><strong>Google releases gemini 3.1 pro preview</strong>: Google just dropped Gemini 3.1 Pro in public preview, touting it as their latest agentic coding and state-of-the-art reasoning model. It scored 77.1% on ARC-AGI-2, a huge jump from its predecessor, and leads most coding benchmarks. This model supports multimodal inputs and offers a 1 million token context window, positioning it as a strong contender in the AI model race, as reported by various sources. <a href="https://app.daily.dev/posts/PjE6Xo1XG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent adoption and impact in software development</strong>: AI coding tools are fundamentally shifting developer roles, with 93% of developers now using AI assistants, moving from writing code to reviewing and orchestrating AI agents. While productivity gains are reported, concerns are rising about AI-generated code introducing subtle production bugs and the ‘AI slop problem’ in open source, where AI-generated PRs create extra review burden for maintainers. This highlights the need for robust engineering maturity and new roles like Agent Orchestration Engineer. <a href="https://app.daily.dev/feed-by-ids?id=037a5YIjx&amp;id=0Q86tGAco&amp;id=0tYkAozTe&amp;id=1v9gMZMqJ&amp;id=2LocixQWJ&amp;id=2WWQ65Wm6&amp;id=2Xamq3Lpe&amp;id=3Kdb5qCcA&amp;id=4CjaWv3pP&amp;id=6KUpHm9AN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>MySQL governance under scrutiny</strong>: Over 248 database professionals, including engineers from major players like Percona and MariaDB, have signed an open letter urging Oracle to transfer MySQL governance to an independent non-profit foundation. They cite declining commit activity, a shrinking contributor pool, lack of AI-era features, and growing fragmentation as key concerns, pushing for a community-led model similar to PostgreSQL to revitalize the database. <a href="https://app.daily.dev/posts/GY7vcPtOu" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Vitest 3.x crushes jest 30 in performance benchmarks</strong>: Vitest 3.x shows significantly faster cold starts (5.6x) and watch mode re-runs (28x) compared to Jest 30 in a 50,000-test monorepo, along with 57% less peak memory usage, according to SitePoint. <a href="https://app.daily.dev/feed-by-ids?id=60rKeST9w&amp;id=9RmovJfdu&amp;id=GsL0A9M1P&amp;id=T3G4VUzdk&amp;id=yRN3HZsMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VectorNest: A new open-source svg editor for web developers</strong>: VectorNest, a free, MIT-licensed, browser-based SVG editor, offers a streamlined workflow for front-end developers creating web icons and UI elements, producing smaller, web-optimized SVG files compared to Illustrator, as highlighted by SitePoint. <a href="https://app.daily.dev/posts/9RmovJfdu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js memory usage halved with node-caged</strong>: Cloudflare, Igalia, and the Node project collaborated on <code>node-caged</code>, a Docker image with V8 pointer compression that can halve Node.js memory usage in production, as reported by Node Weekly. <a href="https://app.daily.dev/posts/Mm7b9KCfF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New cli tool for ai agents to control mobile devices</strong>: Callstack released <code>agent-device</code>, an open-source CLI tool enabling AI agents to programmatically control iOS and Android devices, supporting simulators, emulators, and physical hardware. <a href="https://app.daily.dev/posts/Ekfc0OFB7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apache Kafka 4.2.0 introduces share groups and streams ga</strong>: Apache Kafka 4.2.0 is out with production-ready Share Groups (Kafka Queues) and GA status for Kafka Streams’ server-side rebalance protocol, alongside improved observability and security features, according to Confluent Blog. <a href="https://app.daily.dev/posts/0DcTaIDm9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google cloud spanner now supports cassandra cql apis</strong>: Google Cloud Spanner now offers general availability for native CQL (Cassandra Query Language) API support, allowing existing Apache Cassandra applications to migrate with minimal code changes and leverage Spanner’s global ACID transactions and high availability, as announced by Google Cloud. <a href="https://app.daily.dev/posts/CKaLROkeW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rivian launches apple watch app and software updates</strong>: Rivian is rolling out an Apple Watch companion app for vehicle controls and a broader software update introducing ‘kneel mode’ and a cold weather battery indicator. <a href="https://app.daily.dev/feed-by-ids?id=VXVQM04su&amp;id=WnnenuOfb" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Claude pro integrates with microsoft powerpoint</strong>: Oh, joy! Claude Pro now integrates with Microsoft PowerPoint, allowing users to ‘pull context from their daily tools directly into presentations.’ Because what we really needed was another AI trying to make our slides for us. I’m sure this will lead to groundbreaking, original presentations and not just more AI-generated corporate boilerplate. <a href="https://app.daily.dev/posts/oLQ4V0bPw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Trajectory explorer tool announced for ai agent traces</strong>: Someone announced ‘Trajectory Explorer,’ a new tool for ‘navigating AI agent traces’ and providing a ‘sane way’ to inspect AI agent decisions. Because apparently, AI agents are so complex now, we need another tool just to figure out what the hell they’re doing. ‘First sane way,’ they say? I’ll believe it when I see an AI agent actually make a sane decision. <a href="https://app.daily.dev/posts/AzPQg0qJn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI will solve supply chain failures, apparently</strong>: A new experiment claims Claude Opus 4.6 can diagnose supply chain delivery failures, replacing ‘static dashboards’ for root-cause analysis. So, instead of fixing the actual supply chain issues, we’re just going to throw an AI at it and hope it magically sorts out the mess. Because, you know, a chatbot is definitely going to understand why that container is stuck in Suez better than a human. <a href="https://app.daily.dev/posts/gdtsbiLsr" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
