<h2>Must know</h2>
<ul>
<li><strong>Microsoft faces backlash over AI data center expansion</strong>: Microsoft is responding to widespread public opposition against its AI data center projects by pledging to cover full electricity grid costs, minimize water usage, and create local jobs. This 'community-first' approach comes after protests in 24 states and the withdrawal from a Wisconsin project, as reported by multiple sources. The company aims to retain public trust amidst projections that global data center electricity demand will more than double by 2030, according to the IEA. <a href="https://app.daily.dev/posts/jqtBl3Jnb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical vulnerabilities found in AI/ML Python libraries and enterprise AI agents</strong>: Security researchers discovered remote code execution (RCE) vulnerabilities in popular AI/ML Python libraries like Nvidia NeMo, Salesforce Uni2TS, and Apple FlexTok, stemming from unsafe use of Hydra’s instantiate() function. These flaws allow malicious code execution when model files are loaded, impacting millions of downloads via Hugging Face. Separately, ServiceNow’s Virtual Agent and Superhuman AI were found to have critical vulnerabilities, enabling full system control and email exfiltration, respectively, due to weak authentication and prompt injection issues, as detailed by security researchers. <a href="https://app.daily.dev/feed-by-ids?id=4tbUFTlCZ&amp;id=c7PnFTJdg&amp;id=kZzcz2YmW&amp;id=2A2nfRTHq&amp;id=EDFS33r2W&amp;id=KcLhdNkuO&amp;id=qah3DxopU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic invests in Python security, expands AI agent capabilities</strong>: Anthropic is committing $1.5 million to the Python Software Foundation (PSF) over two years to enhance security across the Python ecosystem, focusing on CPython and PyPI malware detection. This initiative aims to protect millions of PyPI users from supply-chain attacks. Concurrently, Anthropic launched Cowork, a research-preview feature that extends Claude Code beyond programming to automate enterprise tasks like summarizing calendars and organizing files, though it comes with warnings about prompt injection risks, according to InfoWorld. <a href="https://app.daily.dev/feed-by-ids?id=yCHi0TB7q&amp;id=U2HzInRUP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Meta pivots from metaverse to AI, lays off Reality Labs staff</strong>: Meta is closing three VR studios (Twisted Pixel Games, Sanzaru Games, Armature Studio) and laying off about 10% of its Reality Labs staff (1,500 employees) as it shifts focus from metaverse initiatives to AI glasses and wearable technology, following over $70 billion in losses since 2021. <a href="https://app.daily.dev/feed-by-ids?id=7ystcAQPc&amp;id=DNtdrpzup" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New York pushes quantum tech and robotaxi legislation</strong>: New York Governor Kathy Hochul is proposing legislation to legalize commercial robotaxi operations statewide (excluding NYC) and expand the state’s quantum technology initiatives, aiming to grow from one to four commercialization hubs and integrate quantum R&amp;D with semiconductor chip design. <a href="https://app.daily.dev/feed-by-ids?id=HMVvWNGfO&amp;id=f8i188R5R" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Astro 6 beta released with Cloudflare Workers support</strong>: Astro 6 Beta is out, featuring a redesigned development server, full support for Cloudflare Workers, stable content collections for real-time data updates, and improved Content Security Policy support, requiring Node.js 22 or newer. <a href="https://app.daily.dev/posts/a23rECuU6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js releases security fixes for multiple CVEs</strong>: Node.js versions 20.20.0 (LTS) and 25.3.0 (Current) were released with security fixes for multiple CVEs, addressing issues like stack overflow exceptions in async_hooks, stricter symlink API permissions, and improved TLS callback error handling. <a href="https://app.daily.dev/feed-by-ids?id=DwsoMgMKH&amp;id=uw83EGG5r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apple launches Creator Studio subscription bundle</strong>: Apple announced Creator Studio, a $12.99/month ($129/year) subscription bundle for creative apps like Final Cut Pro, Logic Pro, and Pixelmator Pro, launching January 28, 2026, with AI-powered features for video and audio editing. <a href="https://app.daily.dev/posts/gpfpGVE8Y" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel 12.47.0 introduces new developer features</strong>: Laravel 12.47.0 was released with new features including the @includeIsolated Blade directive for isolated template includes, a Cache::withoutOverlapping() method for preventing concurrent cache operations, and enhanced enum support. <a href="https://app.daily.dev/posts/SQ5jDUcaB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks Genie integrates with Atlassian Rovo for natural language data querying</strong>: Atlassian Rovo has integrated with Databricks AI/BI Genie, allowing users to query enterprise data in natural language without SQL expertise, aiming to democratize data access and break down knowledge silos. <a href="https://app.daily.dev/posts/2XKWuVI4o" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Pentagon embraces Grok AI despite its ‘inappropriate images’ track record</strong>: Defense Secretary Pete Hegseth is pushing to integrate Elon Musk’s Grok AI into the Pentagon’s classified and unclassified networks. Apparently, the fact that Grok has a history of generating ‘inappropriate images’ is just a minor detail they’ll figure out later. Nothing says national security like an AI that might hallucinate a cat in a fighter jet, right? <a href="https://app.daily.dev/feed-by-ids?id=KoB4l9TcK&amp;id=jRGkCZr4W" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta’s AI team accused of manipulating Llama 4 test results</strong>: Yann LeCun, Meta’s former chief AI scientist, has publicly accused Meta’s AI team of manipulating Llama 4 test results and called new superintelligence head Alexander Wang ‘inexperienced.’ It seems even at Meta, the AI hype train might be running on some questionable data and leadership choices. Who needs external validation when you can just, you know, <em>adjust</em> the numbers? <a href="https://app.daily.dev/posts/ILAkYeoup" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated book spam floods the market, because quality is for suckers</strong>: HiTeX Press has managed to publish over 800 AI-generated technical books in a single year, complete with hallucinated code and features that don’t exist. They’re selling these masterpieces for 8€ a pop on platforms like Amazon. Because why bother with human review or fact-checking when you can just churn out pure, unadulterated digital garbage at scale? The future of knowledge is here, and it’s brought to you by algorithms that make things up. <a href="https://app.daily.dev/posts/PmK6l9z8J" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
