<h2>Must know</h2>
<ul>
<li><strong>Meta pivots from metaverse to AI, cuts Reality Labs staff</strong>: Meta is closing three VR studios and laying off 10% of its Reality Labs workforce, totaling about 1,500 employees. This move signals a major strategic shift away from the metaverse and VR initiatives, which have lost over $70 billion since 2021, towards AI-powered glasses and wearable technology. CTO Andrew Bosworth has scheduled an urgent meeting to address the changes, as reported by multiple sources. <a href="https://app.daily.dev/posts/7ystcAQPc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft’s AI data center expansion faces backlash</strong>: Microsoft is facing significant community opposition in 24 U.S. states over its AI data center expansion, leading to abandoned projects and ongoing resistance. In response, Microsoft launched a “Community-First AI Infrastructure” initiative, promising to cover all power costs, minimize water usage, and create local jobs. This initiative aims to prevent residential electricity bills from rising due to increased grid demand, a major concern for activists. <a href="https://app.daily.dev/posts/jqtBl3Jnb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare fined by Italy over DNS refusal</strong>: Italy’s communications regulator, AGCOM, has fined Cloudflare €14.2 million for refusing to block pirate sites on its 1.1.1.1 DNS resolver, citing the Piracy Shield law. Cloudflare argues compliance is technically infeasible without increasing latency and blocking legitimate sites, threatening to withdraw its servers from Italy. CEO Matthew Prince is seeking U.S. government support and has notified the International Olympic Committee about potential cybersecurity risks for the 2026 Winter Olympics in Italy. <a href="https://app.daily.dev/posts/Q0IbcQd4P" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Google and Meta expand AI infrastructure and models</strong>: Google released Gemini 3, a multimodal AI platform with a 1M+ token context window, and Veo 3.1 for vertical video creation, while Meta announced Meta Compute, a massive AI infrastructure scale-up aiming for tens of gigawatts of power capacity. <a href="https://app.daily.dev/feed-by-ids?id=2S31ZjKRT&amp;id=NEGZugNBg&amp;id=kwRMqebwB&amp;id=qjYQj2ENJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic commits $1.5 million to Python Software Foundation</strong>: Anthropic has pledged $1.5 million over two years to the Python Software Foundation (PSF) to enhance Python ecosystem security, focusing on automated malware detection for PyPI packages and supporting CPython core development. <a href="https://app.daily.dev/posts/yCHi0TB7q" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Slackbot relaunches as AI agent for enterprises</strong>: Salesforce has redesigned Slackbot as an AI agent for Business+ and Enterprise+ customers, leveraging context from Slack and third-party tools to answer questions, organize work, and schedule meetings. <a href="https://app.daily.dev/posts/VpXLIsjSp" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quantum computing sees new products and investments</strong>: The quantum computing industry is buzzing with activity, including SuperQ’s ChatQLM launch, Xanadu’s leadership changes, Penn State’s security vulnerability findings, and significant investments in quantum firms, alongside IBM’s Nighthawk processor release. <a href="https://app.daily.dev/feed-by-ids?id=EJu78wwgZ&amp;id=J8FTmD4cu&amp;id=JmIuGeLwy&amp;id=Lx6jVDI9P&amp;id=MkzC4PEYB&amp;id=QPb7M1IyK&amp;id=RHyOC3Vzq&amp;id=RONvWGnCv&amp;id=dtnIdUSBt&amp;id=hhNibwkwa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Deepgram raises $130M, OpenAI acquires Torch</strong>: Voice AI company Deepgram secured $130 million in Series C funding at a $1.3 billion valuation and acquired Ofone, while OpenAI acquired health records AI startup Torch for an estimated $100 million to integrate into ChatGPT Health. <a href="https://app.daily.dev/feed-by-ids?id=ZgbOyOa65&amp;id=cfKcRxGmX&amp;id=qKOwUFbrZ&amp;id=qxMWIc2TX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js releases security updates</strong>: Node.js 20.20.0 (LTS) and 25.3.0 (Current) were released, addressing multiple security vulnerabilities, including disabling futimes with the permission model enabled and tightening symlink API permissions. <a href="https://app.daily.dev/feed-by-ids?id=DwsoMgMKH&amp;id=Wy4br4Zyw&amp;id=Zr0Dv9dwZ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WordPress 7.0 to drop PHP 7.2/7.3 support</strong>: The WordPress core team is preparing for the 7.0 release, which will discontinue support for PHP 7.2 and 7.3, and is seeking volunteers for the upcoming version. <a href="https://app.daily.dev/posts/xho5u0zys" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents are unreliable and insecure, says Signal</strong>: Signal’s leadership, Meredith Whittaker and VP, are out here stating the obvious: agentic AI systems are inherently unreliable and insecure. They point to Microsoft’s Recall as a prime example of user activity being stored in vulnerable databases. Apparently, these AI agents are ‘probabilistic, not deterministic,’ meaning they fail spectacularly at multi-step tasks. Who knew that relying on a glorified autocomplete for critical functions might be a bad idea? <a href="https://app.daily.dev/feed-by-ids?id=TJt6uaOsV&amp;id=XnmeLgefE&amp;id=rjeAcjlYe" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gartner predicts 40% of AI agent projects will fail</strong>: Gartner, in its infinite wisdom, predicts that 40% of agentic AI projects will be canceled by 2027. The reason? Not the AI models themselves, but a shocking lack of ‘guardrails.’ Apparently, treating AI agents like human employees with strict governance, technical constraints, and rigorous testing is too much to ask. Who would’ve thought that letting an AI chatbot run wild could lead to costly errors, like that Chevrolet bot trying to sell a car for a dollar? <a href="https://app.daily.dev/feed-by-ids?id=9X1ifJmCu&amp;id=QnCK3ryB4&amp;id=SVMT3qR0j" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open source projects are getting screwed by AI</strong>: Tailwind CSS saw an 80% revenue drop because AI models are just scraping their documentation, effectively killing their monetization. Now, open-source maintainers are scrambling, switching to stricter licenses like AGPL-3.0 to stop AI from feasting on their code without compensation. And if that wasn’t enough, they’re drowning in low-quality, AI-generated contributions. It seems the ‘open’ in open source is becoming a liability when everyone’s trying to build their next billion-dollar AI on your free labor. <a href="https://app.daily.dev/feed-by-ids?id=2a8HRxzCb&amp;id=Ml47Y2Hd6&amp;id=g50WJmopi" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
