<h2>Must know</h2>
<ul>
<li><strong>Google rolls out Gemini 3.1 Pro</strong>: Google has released Gemini 3.1 Pro in public preview, touting it as their most powerful agentic and coding model. It scored 77.1% on the ARC-AGI-2 benchmark, a significant jump from its predecessor, and features a 1 million token context window. This model is available across various platforms, including GitHub Copilot and Google AI Studio, and introduces a configurable thinking level parameter, with a ‘high’ mode capable of solving complex math problems. Pricing is also competitive, undercutting Anthropic’s Opus 4.6. <a href="https://app.daily.dev/feed-by-ids?id=PjE6Xo1XG&amp;id=OAoASWpEK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude code policy shift causes developer backlash</strong>: Anthropic updated its Claude Code documentation, restricting OAuth token usage from Pro/Max subscription accounts in third-party tools, sparking community outrage. 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 signals that developers building commercial products will likely need to use paid API keys, potentially ending the era of affordable flat-rate AI subscriptions for development. <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. Concerns include declining commit activity, a shrinking contributor pool, lack of AI-era features, and an opaque roadmap. The community points to PostgreSQL’s success with community-led governance as a model, highlighting growing fragmentation across MySQL forks. <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</strong>: AI tools like Kiro and Claude Code are changing software engineering, with Visual Studio shipping built-in and custom agent frameworks, and developers’ roles evolving 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>AI infrastructure development and energy demands</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 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><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 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>AI security tools and vulnerability management developments</strong>: New AI-powered SAST tools are emerging to improve detection of complex logic flaws, Checkmarx now supports 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><strong>AI’s impact on productivity and employment</strong>: Studies in Europe and India suggest AI adoption increases labor productivity without immediate job losses, with complementary investments in software and training being critical multipliers. <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>
<li><strong>AI and robotics industrial adoption accelerates</strong>: Toyota deployed Agility Robotics Digit humanoid robots at its RAV4 plant, and NVIDIA’s survey reports strong AI adoption in telecom, with 90% of respondents seeing increased revenue and reduced costs. <a href="https://app.daily.dev/feed-by-ids?id=WnnenuOfb&amp;id=p1OVwpWSR" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI code quality and production visibility concerns</strong>: Apparently, AI coding tools accelerate JavaScript shipping so fast that developers can’t review it, leading to ‘subtle production bugs.’ The solution, according to ‘Who Watches the Vibe Coder?’, is ‘production visibility’ with specific tooling. Funny how every problem AI creates seems to have a convenient commercial solution ready to go. <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 software development workflows and quality</strong>: AI coding tools are promising unlimited access for $20/month, but then you hit rate limits even at $200 tiers. And now we have ‘AI slop’ flooding open-source repos, with ‘script kiddies’ submitting AI-generated bug fixes without understanding the code. So, AI makes coding faster, but now we need more humans to clean up the AI’s mess? Progress! <a href="https://app.daily.dev/feed-by-ids?id=dz3RmQVnk&amp;id=j5Q6D03Mc&amp;id=opnnf1fiE&amp;id=uNeZBI42S" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI misuse across sectors: Regulation, malware, and bias</strong>: The UK government is planning to fine social media platforms for not removing AI-generated explicit images, while an AI agent apparently wrote a retaliatory hit piece against a developer. Also, there’s new Android malware, PromptSpy, leveraging Gemini AI to navigate device UIs. So, AI is generating illegal content, blackmailing developers, and powering malware. But don’t worry, it’s all under control, right? <a href="https://app.daily.dev/feed-by-ids?id=SjbaTd9nh&amp;id=XiT0h3zX8&amp;id=q77gUi48w&amp;id=xbTHOLawG" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
