<h2>Must know</h2>
<ul>
<li><strong>MySQL community demands independent governance from Oracle</strong>: A coalition of over 248 database professionals from companies like Percona and MariaDB has issued an open letter urging Oracle to transfer MySQL’s governance to an independent non-profit foundation. They cite declining commit activity, a shrinking contributor base, lack of AI-era features, and an opaque roadmap under Oracle’s stewardship. The community proposes a foundation model with a technical steering committee, aiming for a PostgreSQL-like success story, though analysts are skeptical about power dynamics if Oracle retains trademark control. <a href="https://app.daily.dev/posts/GY7vcPtOu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google releases Gemini 3.1 Pro with massive context window</strong>: Google has launched Gemini 3.1 Pro in public preview, touting it as their most powerful agentic and coding model yet. It features a 1 million token input context window and a 64k output token limit, scoring 77.1% on the ARC-AGI-2 benchmark. The model is available across various Google and GitHub platforms, with pricing set at $2 per million input tokens and $12 per million output tokens, undercutting competitors like Anthropic’s Opus 4.6. <a href="https://app.daily.dev/posts/PjE6Xo1XG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude Code policy shift causes developer uproar</strong>: Anthropic updated its Claude Code documentation, restricting OAuth token usage from Pro/Max subscription accounts in third-party tools, leading to community backlash. While Anthropic claims it’s a documentation cleanup, actions like a cease-and-desist to OpenClaw and banning OpenCode from OAuth access suggest a move towards requiring paid API keys for commercial development. This shift raises concerns about the end of affordable flat-rate AI subscriptions for developers building businesses. <a href="https://app.daily.dev/feed-by-ids?id=3lsZ5Vula&amp;id=FQxtQgg9R&amp;id=GYB8zPRsi" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Rust and WebAssembly power terminal recording evolution</strong>: Asciinema, the terminal recording tool, has evolved significantly over 14 years, moving from Python to a multi-language ecosystem including Go, Rust, and Elixir, with its core player now using Rust compiled to WebAssembly for improved performance and a recent CLI recorder rewrite to Rust for live streaming. <a href="https://app.daily.dev/posts/Hk6RIbomY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents reshape software development and security</strong>: Agentic AI is transforming software development, with tools like Kiro and Claude Code boosting productivity, while new solutions like Link11’s dashboard and Azazel’s eBPF tracer emerge to manage and secure AI crawler traffic and agent actions at the kernel level. <a href="https://app.daily.dev/feed-by-ids?id=zxo3wKZNk&amp;id=037a5YIjx&amp;id=4CjaWv3pP&amp;id=7XxAvY7Df&amp;id=Bl6akHr3q&amp;id=DqRWRtC6t&amp;id=F8eGl7g0G&amp;id=LCg5ueVO7&amp;id=SBASpAo7A&amp;id=UJqBwwcg1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hugging Face launches community evals for AI model benchmarking</strong>: Hugging Face introduced “Community Evals,” a beta feature aimed at transparent, community-driven AI model evaluation, allowing benchmark datasets to host leaderboards and aggregate results to address inconsistencies in reported scores. <a href="https://app.daily.dev/posts/9WqEmBh9G" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI memory and agent management solutions emerge</strong>: New solutions like Mengram’s open-source AI memory API and Reload’s Epic platform are addressing AI agent memory limitations, with Mengram focusing on semantic, episodic, and procedural memory types, and Reload providing shared long-term memory for AI coding teams. <a href="https://app.daily.dev/feed-by-ids?id=0tYkAozTe&amp;id=MQxvV1b9u" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI infrastructure demands drive power plant and data center expansion</strong>: The surging demand for AI infrastructure is leading to massive investments, including SoftBank’s plan for a $33 billion natural gas power plant and Cisco’s Secure AI Factory with NVIDIA, highlighting the significant energy and hardware requirements for AI development. <a href="https://app.daily.dev/feed-by-ids?id=IYsqAR4Cy&amp;id=VBQEBjah6&amp;id=1mD3jO5pQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apache Polaris graduates to top-level project</strong>: Apache Polaris, an open-source catalog for Iceberg and data lakehouse interoperability originally from Snowflake, has graduated to a Top-Level Project under the Apache Software Foundation, signaling mature governance and momentum for open data lakehouse standards. <a href="https://app.daily.dev/posts/ovRl52TUb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale workload identity federation now generally available</strong>: Tailscale’s workload identity federation is now generally available, enabling CI/CD pipelines and cloud workloads to authenticate using short-lived OIDC tokens instead of static credentials, with full CRUD support via API, Terraform integration, and automatic cloud token discovery. <a href="https://app.daily.dev/posts/ekx9IIs1v" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI tools are saving people time and money, apparently</strong>: Matt Wolfe, a self-proclaimed expert, claims AI tools like Perplexity, Claude, and Cursor reduce his daily work to about two hours. While it’s great he’s found his personal AI nirvana, this sounds suspiciously like the kind of anecdotal ‘productivity porn’ that ignores the actual complexities of real-world engineering. Maybe he’s just really good at delegating to his digital assistants, or maybe his ‘work’ is just generating more content about how AI saves him time. <a href="https://app.daily.dev/posts/X2M941duB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI code quality concerns: The new ‘vibe coding’</strong>: According to ‘Who Watches the Vibe Coder?’, AI coding tools are accelerating JavaScript shipping faster than developers can review it, leading to subtle production bugs. The article then conveniently advocates for ‘production visibility’ tools like TrackJS as the solution. So, AI makes code faster but buggier, and the fix is to buy more tools to monitor the bugs. Sounds like a perfect ecosystem for vendors, less so for actual code quality. <a href="https://app.daily.dev/feed-by-ids?id=2Xamq3Lpe&amp;id=IR0A3b0XK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are the new microservices, but with more buzzwords</strong>: A new tool called ‘Trajectory Explorer’ is being touted as the ‘first sane way to navigate agent traces,’ whatever that means. Meanwhile, another article satirizes the endless parade of software architecture buzzwords and AI hype. It seems we’re just replacing one set of over-engineered, poorly understood concepts with another, now with added ‘agent traces’ and ‘trajectory exploration.’ Because nothing says pragmatic software design like needing a new tool just to understand what your AI agent is doing. <a href="https://app.daily.dev/feed-by-ids?id=AzPQg0qJn&amp;id=SwJH4RR0B" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
