<h2>Must know</h2>
<ul>
<li>MySQL governance under Oracle challenged by community: Over 248 database professionals, including engineers from 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 an opaque roadmap as reasons for concern. The community proposes a foundation model with a technical steering committee, while Oracle would retain commercial offerings and trademarks, a setup analysts are skeptical will fully resolve power dynamics. <a href="https://app.daily.dev/posts/GY7vcPtOu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Anthropic’s Claude code policy shift causes developer backlash: 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 just a documentation cleanup, official legal documents explicitly state these tokens are for Claude Code and <a href="http://Claude.ai" target="_blank" rel="noopener nofollow">Claude.ai</a> only. This move, coupled with cease-and-desist letters to developers, suggests the era of affordable flat-rate AI subscriptions for commercial development might be ending, pushing developers towards more expensive API key authentication. <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>
<li>OpenAI’s persona identity verification system raises surveillance concerns: Security researchers uncovered that Persona, OpenAI’s KYC provider, operates a watchlist screening infrastructure that files reports to federal agencies. The system, exposed via unprotected source maps, includes biometric face databases, 269 verification checks, and crypto address monitoring. This raises serious questions about BIPA compliance, data retention, and the lack of transparency or appeal mechanisms for denied users, especially given its use by government entities. <a href="https://app.daily.dev/posts/JdSs9FMH3" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li>Agentic AI reshaping software development: AI tools like Kiro and Claude Code are shifting the productivity baseline for software engineers, with Visual Studio shipping built-in and custom agent frameworks, and McKinsey data indicating 23% of organizations are scaling agentic AI. <a href="https://app.daily.dev/feed-by-ids?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&amp;id=Uw1m3wdti" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI memory and agent management solutions emerge: Mengram launched an open-source AI memory API with semantic, episodic, and procedural memory types, while Reload raised $2.275 million for Epic, a platform providing shared long-term memory for AI coding agents. <a href="https://app.daily.dev/feed-by-ids?id=0tYkAozTe&amp;id=MQxvV1b9u" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI infrastructure development and energy demands: Cisco announced the Secure AI Factory with NVIDIA, a modular architecture for AI compute, networking, storage, and security, while SoftBank plans a $33 billion, 9.2 gigawatt natural gas power plant to meet surging AI data center energy demands. <a href="https://app.daily.dev/feed-by-ids?id=IYsqAR4Cy&amp;id=VBQEBjah6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI agent and model updates across GitHub, Heroku, and AWS: GitHub Copilot now integrates code referencing, Heroku launched a new standard plan for its Managed Inference and Agents add-on with new models, and AWS announced Aurora DSQL integration with Kiro powers and AI agent skills. <a href="https://app.daily.dev/feed-by-ids?id=kcB1LfSD2&amp;id=lOJYvaow4&amp;id=rsFWrvtAL&amp;id=vDLyK99UX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Global AI expansion, data sovereignty, and market rivalries: OpenAI partnered with India’s Tata Group for 100MW of AI-ready data center capacity and plans new offices, while Google Cloud emphasizes multicloud freedom and interoperability for Europe’s AI future, and NTT DATA uses synthetic data to boost Japanese AI development. <a href="https://app.daily.dev/feed-by-ids?id=6fxZc9j6p&amp;id=KUvvYNNG9&amp;id=RW47wQGgP&amp;id=XnTMjJfPz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI security tools and vulnerability management developments: AI-powered SAST tools are emerging to improve detection and triage of complex logic flaws, Checkmarx announced support for AWS’s Kiro AI coding tool, and BARBIE is a new backdoor detection method for deep learning models. <a href="https://app.daily.dev/feed-by-ids?id=2TCivC1ph&amp;id=9x8XnEM3d&amp;id=CBTU3pEb0&amp;id=zsAWpz3GM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Workload identity federation now generally available: 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 API keys or secrets, with full CRUD support via API and Terraform integration. <a href="https://app.daily.dev/posts/ekx9IIs1v" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li>AI productivity experiment: developers feel productive, but are they?: Wix Engineering’s 5-week experiment found that while developers felt more productive using tools like Cursor, actual measurable output gains were minimal. Claude Code showed significant gains, but only after developers stopped micromanaging it. This ‘AI Curse’ means individual productivity explodes, but organizational coordination and review processes become new bottlenecks. So, you might feel like a superhero, but your team’s still stuck in traffic. <a href="https://app.daily.dev/posts/qQhctEUPB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI code quality and production visibility concerns: Apparently, AI coding tools accelerate JavaScript shipping so fast that developers can’t review it, leading to ‘subtle production bugs’ like missing guards and async race conditions. The solution? ‘Production visibility’ tools like TrackJS, which, coincidentally, the article explicitly advocates for. It’s almost like the problem was invented to sell the solution. Who watches the vibe coder? The one selling the monitoring tools, apparently. <a href="https://app.daily.dev/feed-by-ids?id=2Xamq3Lpe&amp;id=IR0A3b0XK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI’s impact on work: cognitive debt and authenticity: A historian warns of ‘cognitive debt’ from outsourcing thinking to AI, arguing that writing is an irreplaceable form of public thought. Meanwhile, CEOs of Read AI and Lucidya claim AI replaces tasks, not roles, citing their five-person support team serving millions. So, either AI is making us dumber and less authentic, or it’s enabling lean teams to serve millions. Pick your poison, but don’t expect AI to write your next great novel, unless it’s a marketing one. <a href="https://app.daily.dev/feed-by-ids?id=4HNvjLVYH&amp;id=QMBkqUnaK" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>