Changelog

Developer Voices

Google Cloud

Hacker News

Hugging Face

InfoWorld

<h2>Must know</h2>
<ul>
<li><strong>Google rolls out 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 boasts a 1 million token input context window and scored 77.1% on the ARC-AGI-2 benchmark, a significant jump from its predecessor. This model is available across various platforms, including GitHub Copilot, Google AI Studio, and Vertex AI, and is priced competitively against rivals 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, sparking community backlash. While Anthropic claims it’s just a documentation cleanup, actions like sending a cease-and-desist to OpenClaw and banning OpenCode from OAuth access suggest a shift. This move raises concerns that 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><strong>MySQL governance under Oracle challenged by community</strong>: 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 key concerns. The community points to PostgreSQL’s success with community-led governance as a model, though analysts are skeptical about resolving power dynamics if Oracle retains trademark control. <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>AI agents reshape software development workflows</strong>: AI tools like Kiro and Claude Code are shifting the productivity baseline for software engineers, with Visual Studio shipping built-in preset agents and a custom agent framework, while developers’ roles evolve into Agent Orchestration Engineer and LLMOps Specialist. <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><strong>AI memory and agent management solutions emerge</strong>: Mengram launched an open-source AI memory API with semantic, episodic, and procedural memory types, while Reload introduced Epic, an AI workforce management 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><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 infrastructure development and energy demands surge</strong>: Cisco announced the Secure AI Factory with NVIDIA, a modular architecture for AI compute, networking, storage, and security, while SoftBank plans a massive 9.2 gigawatt natural gas power plant to meet AI data center energy needs. <a href="https://app.daily.dev/feed-by-ids?id=IYsqAR4Cy&amp;id=VBQEBjah6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI and robotics industrial adoption accelerates</strong>: Toyota deployed seven Agility Robotics Digit humanoid robots at its RAV4 plant for auto parts unloading, and NVIDIA’s survey reports 90% of telecom respondents see AI increasing revenue and reducing costs, with 89% planning increased AI spending. <a href="https://app.daily.dev/feed-by-ids?id=WnnenuOfb&amp;id=p1OVwpWSR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agent and model updates across GitHub, Heroku, and AWS</strong>: 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 Amazon 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><strong>Redwood Materials expands energy storage for AI data centers</strong>: Redwood Materials, founded by former Tesla CTO JB Straubel, is rapidly expanding its energy storage business unit, repurposing EV batteries to meet the surging power demands of AI data centers, securing $425 million in funding from Google and Nvidia. <a href="https://app.daily.dev/posts/1mD3jO5pQ" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Wix engineering’s AI productivity experiment: feelings vs facts</strong>: Wix Engineering ran a 5-week experiment on AI coding tools, finding that while developers <em>felt</em> more productive with Cursor, there was “almost no measurable productivity lift.” Claude Code, however, did show significant gains once developers stopped micromanaging it. So, your feelings are valid, but maybe check the GitHub data before you declare yourself a 10x engineer. <a href="https://app.daily.dev/posts/qQhctEUPB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI code quality and production visibility concerns</strong>: Apparently, AI coding tools are so fast at shipping JavaScript that developers can’t review it all, leading to “subtle production bugs like missing guards, async race conditions, invisible coupling, and swallowed exceptions.” The solution? “Production visibility” tools like TrackJS. Because nothing says progress like needing another tool to clean up the mess made by the last tool. <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’s impact on productivity and employment: the usual suspects</strong>: A European study claims AI boosts labor productivity by 4% with no short-term job losses, while an Indian think tank says AI isn’t an immediate threat to their IT sector. Meanwhile, some Silicon Valley types are still arguing that fears about AI destroying jobs are just “historical mistakes.” It’s almost like everyone has a vested interest in painting AI as a job creator, right up until it’s time to cut headcount. <a href="https://app.daily.dev/feed-by-ids?id=iZlTkI8Me&amp;id=tiBubXcsb&amp;id=yOsJJlumD" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
