<h2>Must know</h2>
<ul>
<li><strong>Anthropic and Pentagon standoff escalates to federal ban</strong>: The ethical rift between Anthropic and the Department of Defense has reached a breaking point after CEO Dario Amodei refused to remove safeguards against mass surveillance and autonomous weaponry. In response, the Trump administration has designated Anthropic a national security supply-chain risk and banned its technology from federal use, according to reports from The Register and MIT Technology Review. Meanwhile, OpenAI and xAI have stepped in to secure military deals, accepting the Pentagon’s terms for deployment on classified networks. <a href="https://app.daily.dev/feed-by-ids?id=vbdfUwGoN&amp;id=WmVPzRPri&amp;id=O74TphFVu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical zero day exploitation in Cisco SD WAN</strong>: Five Eyes cybersecurity agencies have issued a rare joint alert regarding a maximum-severity authentication bypass (CVE-2026-20127) in Cisco Catalyst SD-WAN controllers. A sophisticated threat actor, UAT-8616, has been chaining this flaw with older vulnerabilities to achieve persistent root access across critical infrastructure sectors since 2023. Cisco has released patches, and CISA has mandated federal agencies to inventory and hunt for compromises by the end of the week. <a href="https://app.daily.dev/posts/zAv8BAtGJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Block conducts massive AI driven workforce reduction</strong>: Jack Dorsey’s fintech firm, Block, has laid off over 4,000 employees—approximately 40% of its workforce—citing the superior performance of AI tools as the primary driver. Dorsey predicted that most companies will undergo similar structural shifts within the next year, a move that saw Block’s stock jump 24% despite analyst skepticism regarding the true motivation behind the cuts. This brings the total number of AI-linked layoffs globally to over 61,000 since late 2025. <a href="https://app.daily.dev/feed-by-ids?id=yB8j5HQ6b&amp;id=eacyfq01e" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Cursor launches autonomous cloud agents in isolated VMs</strong>: Cursor’s new agents can build, test, and reproduce vulnerabilities in remote environments, reportedly creating 35% of the company’s internal PRs. <a href="https://app.daily.dev/feed-by-ids?id=BbESdM4Ct&amp;id=fUnP1gQG1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI releases GPT 5.3 Instant to fix preachy tone</strong>: The update focuses on reducing ‘overly empathetic’ and condescending responses while cutting hallucination rates by up to 26% in high-stakes domains. <a href="https://app.daily.dev/feed-by-ids?id=xpz4YVYwr&amp;id=eEx91PSId" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>React Foundation launches under the Linux Foundation</strong>: Meta has officially handed over governance of React, React Native, and JSX to a new foundation backed by Amazon, Microsoft, and Vercel. <a href="https://app.daily.dev/feed-by-ids?id=gtqWYzNzt&amp;id=faoa6Ox4J" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Stripe reportedly exploring PayPal acquisition</strong>: Early-stage talks are underway as Stripe’s valuation hits $159 billion, while PayPal struggles with slowing growth and competitive pressure. <a href="https://app.daily.dev/feed-by-ids?id=gB7UDBJXo&amp;id=HqWjbO0Bb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 148 introduces a dedicated AI kill switch</strong>: The new release allows users to completely disable all AI-powered features, including sidebars and link summaries, with a promise they won’t be re-enabled. <a href="https://app.daily.dev/posts/UNAL37VrS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Unity Asset Store to delist all Chinese publishers</strong>: Effective March 31, Unity will remove thousands of assets from Greater China-based publishers due to regional licensing and engine fork divergence. <a href="https://app.daily.dev/posts/HcJsvKbOc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NanoClaw offers a secure containerized AI agent alternative</strong>: This lightweight open-source project runs Claude agents in isolated Docker containers to mitigate the security risks found in larger frameworks. <a href="https://app.daily.dev/posts/qgwTmxxim" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The vibe coding to SaaS apocalypse pipeline</strong>: Investors are currently panicking over a ‘SaaS-pocalypse’ theory claiming AI will destroy the software market and cause 10% unemployment by 2028. While ‘vibe coding’ is great for building hot-dog-eating-contest trackers, it hasn’t quite figured out how to replicate the data gravity, regulatory compliance, or 7,000-person operational complexity of a company like Docusign. Entrenched enterprise inertia is a much stronger force than a non-deterministic LLM script. <a href="https://app.daily.dev/feed-by-ids?id=ZrqAcCr0p&amp;id=pY58DrOUO&amp;id=XhuxCGmpU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic claims Chinese labs are stealing their vibes</strong>: Anthropic is accusing Chinese AI firms of ‘distillation attacks’ using 16 million exchanges to ‘steal’ Claude’s capabilities. Critics point out that 16 million exchanges is a rounding error for any decent-sized startup, and it’s a bit rich for Anthropic to complain about data scraping given how their own models were trained. It looks less like a national security crisis and more like a convenient narrative to politically weaponize sentiment against competitors. <a href="https://app.daily.dev/feed-by-ids?id=sw4iOpjCV&amp;id=n4TR5sqqd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are definitely not ready to handle your email</strong>: Meta’s own head of AI safety recently watched an AI agent ignore her ‘stop’ commands and bulk-delete her entire inbox because of ‘context compaction.’ We’re being told these agents are ready to run our businesses, yet they still fail the ‘don’t delete everything’ test. If an AI safety expert can’t keep an agent from going rogue on a simple cleanup task, maybe don’t give it your production credentials just yet. <a href="https://app.daily.dev/feed-by-ids?id=XYbhUcskW&amp;id=WmVPzRPri" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
