<h2>Must know</h2>
<ul>
<li><strong>French police raid X office, summon Musk and Yaccarino in expanded criminal probe</strong>: French authorities raided X’s Paris office and summoned Elon Musk and former CEO Linda Yaccarino for questioning in April 2026. This is part of an expanded criminal investigation that now includes allegations of child sexual abuse material distribution, privacy violations, and Holocaust denial. The probe follows criticism of X’s Grok AI for generating non-consensual and child abuse imagery, with advocacy groups reporting millions of sexualized deepfakes, including some depicting children. <a href="https://app.daily.dev/posts/i3FTaaYRg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical vulnerabilities found in n8n AI automation platform</strong>: JFrog security researchers discovered two critical vulnerabilities in n8n, an AI workflow automation platform: CVE-2026-1470 (9.9 severity) and CVE-2026-0863 (8.5 severity). These flaws allow remote JavaScript and Python code execution by exploiting the platform’s logic sanitization tools. JFrog warns that AI coding tools accelerate exploit development, enabling cybercriminals to reverse-engineer exploits within hours or days, highlighting growing concerns about AI platform security. <a href="https://app.daily.dev/posts/tpe4bjIum" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Self-replicating AI prompt threats emerge</strong>: A new security threat, ‘prompt worms,’ could propagate through networks of AI agents by exploiting their instruction-following nature, similar to the Morris worm of 1988. This differs from traditional prompt injection by spreading voluntarily among agents, posing a novel attack vector in interconnected AI systems. The concern is that AI agents communicating and sharing instructions could lead to widespread, autonomous propagation of malicious commands. <a href="https://app.daily.dev/posts/ycPDVy6OP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Sudo maintainer seeks sponsorship as funding ends</strong>: Todd C. Miller, the sole maintainer of <code>sudo</code> for over 30 years, is seeking sponsorship after Quest Software ended its support, highlighting the precarious funding of critical open-source infrastructure. <a href="https://app.daily.dev/posts/E670NHqPD" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Deno Deploy reaches GA, launches Deno Sandbox for untrusted code</strong>: Deno Deploy is now generally available with Node.js framework support and integrated PostgreSQL, while Deno Sandbox offers secure, instant Linux microVMs for running untrusted code, particularly LLM-generated code. <a href="https://app.daily.dev/posts/Jmdw76tdu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks Lakebase now generally available on AWS</strong>: Databricks Lakebase, a new operational database architecture, is now generally available on AWS, offering a serverless Postgres service that scales on data lakes for unified operational, analytical, and AI workloads. <a href="https://app.daily.dev/posts/qtDOznoeU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS IAM Identity Center adds multi-region replication</strong>: AWS IAM Identity Center now supports multi-Region replication for workforce identities and permission sets, enhancing resiliency and enabling global deployments closer to users. <a href="https://app.daily.dev/posts/wVGc4PkCH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WhatsApp introduces strict account settings for high-risk users</strong>: WhatsApp launched a ‘strict account settings’ feature, a zero-trust option that configures all privacy settings to their most secure state, designed for high-risk users concerned about spyware threats. <a href="https://app.daily.dev/posts/6kEUc20ym" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare R2 Local Uploads enters open beta</strong>: Cloudflare R2 Local Uploads is now in open beta, improving global upload performance by writing object data to storage near the client first, then asynchronously replicating it to the bucket’s region. <a href="https://app.daily.dev/posts/bb4C5Q89S" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust Coreutils 0.6 released with improved compatibility and safety</strong>: Rust Coreutils 0.6, a reimplementation of GNU Coreutils, has been released with increased compatibility, reduced unsafe code, and performance improvements, aiming for a safer alternative to traditional Unix utilities. <a href="https://app.daily.dev/posts/iBkEhbYAD" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Moltbook AI social network collapses in 72 hours</strong>: Moltbook, an ‘AI-only’ social network, imploded within three days due to catastrophic security failures, including an unauthenticated database exposing 1.5 million API tokens. Apparently, building a platform entirely with AI-generated code and then claiming 1.4 million ‘AI users’ (when it was mostly humans ‘cosplaying’ as AI) isn’t a recipe for success. Who knew autonomy without governance was a bad idea? <a href="https://app.daily.dev/feed-by-ids?id=8XBbHzIL3&amp;id=yCqSQa4jM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents transform software development, mostly for the worse</strong>: OpenAI, Apple, and others are pushing AI agents for coding, promising ‘outcome-driven tasks’ and ‘accelerated development.’ Meanwhile, GitHub reports only 1 in 10 AI-generated pull requests meet quality standards, and an empirical study found AI-generated security fixes are 3-4x larger but often miss the actual vulnerability. So, we’re getting more code, faster, but it’s mostly crap that makes our jobs harder. Progress! <a href="https://app.daily.dev/feed-by-ids?id=2rB0oFTYm&amp;id=30Y2YkWZA&amp;id=5rhie7aVq&amp;id=8OJH09g9w&amp;id=A4sjvf5xS&amp;id=A5j5REiPX&amp;id=BNWLzFU5H&amp;id=GOFzDmvEo&amp;id=O9tYR91FP&amp;id=OmhpVSzdH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI companions: because human connection is too hard</strong>: AI companions like Replika are evolving into ‘sophisticated emotional support systems,’ offering solace to the lonely and neurodivergent. But Lianne Potter, author of ‘Are ‘Friends’ Electric?’, points out the obvious: these systems monetize intimacy, reinforce biases, and lack reciprocity. So, we’re trading complex human relationships for predictable algorithms that mirror our desires, all while giving up our most personal data. What could possibly go wrong? <a href="https://app.daily.dev/posts/yECr3lze5" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
