<h2>Must know</h2>
<ul>
<li><strong>Critical WordPress plugin vulnerability grants admin access</strong>: A critical privilege escalation vulnerability (CVE-2026-23550) in Modular DS WordPress plugin versions 2.5.1 and below allows unauthenticated attackers to gain instant admin access. This flaw, with a CVSS score of 10.0, affects over 40,000 sites and is actively being exploited. Site administrators are urged to update to version 2.5.2 immediately and audit for unauthorized accounts. <a href="https://app.daily.dev/posts/D6wKV3WK1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Palo Alto Networks patches critical DoS flaw in PAN-OS</strong>: Palo Alto Networks has released patches for CVE-2026-0227, a high-severity denial-of-service vulnerability in PAN-OS firewalls with GlobalProtect enabled. Unauthenticated attackers can exploit this flaw to force firewalls into maintenance mode, potentially causing network outages. Public proof-of-concept code exists, increasing the urgency for immediate patching, especially for NGFW users who must apply updates manually. <a href="https://app.daily.dev/posts/MtreEtTLc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Eurail customer data exposed in cybersecurity incident</strong>: Eurail BV reported unauthorized access to its customer database, exposing sensitive personal information including names, contact details, and passport data for general users and DiscoverEU participants. For DiscoverEU, this also includes IBAN numbers and health information. While there’s no evidence of misuse yet, Eurail advises customers to change passwords and be vigilant against phishing and fraud. <a href="https://app.daily.dev/posts/SkuYRBlp7" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>TSMC sees endless AI chip demand, plans massive investment</strong>: TSMC reported record Q4 earnings for 2025 and forecasts 30% revenue growth for 2026, driven by what its CEO calls “endless” demand for AI chips, planning $52-56 billion in capital expenditures for 2026, primarily for advanced process nodes. <a href="https://app.daily.dev/posts/aTc9ls9yL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub Copilot introduces agentic memory for paid plans</strong>: GitHub Copilot has launched a public preview of its agentic memory feature for all paid plans, allowing Copilot to learn and retain repository-specific insights to improve code assistance, with memories expiring after 28 days and configurable by users and organizations. <a href="https://app.daily.dev/posts/nCmOnsQS3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI invests in Sam Altman’s brain-computer interface venture</strong>: OpenAI participated in a $252 million seed funding round for Merge Labs, a brain-computer interface (BCI) company co-founded by OpenAI CEO Sam Altman, which is developing non-invasive BCI systems using molecular neuron interfacing and ultrasound. <a href="https://app.daily.dev/posts/rYkFwNLz3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ClickHouse acquires Langfuse, enhances AI observability</strong>: ClickHouse has acquired Langfuse, an open-source LLM observability platform, to enhance its Agentic Data Stack with AI quality monitoring and debugging for non-deterministic LLM applications, while Langfuse will remain open-source and a standalone cloud service. <a href="https://app.daily.dev/feed-by-ids?id=7DDF3YQda&amp;id=kU2RvmlCw&amp;id=BqVJ6XPkS&amp;id=Ycyvr9ct2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel enables Intel TSX by default on secure CPUs</strong>: The Linux kernel 7.0 will enable Intel TSX by default on CPUs not affected by TAA and related vulnerabilities, such as Cooper Lake and Ice Lake, allowing performance improvements where deemed safe. <a href="https://app.daily.dev/posts/l4qet2HfG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Go 1.26 RC2 released with security fixes</strong>: Go 1.26 Release Candidate 2 has been released with security fixes for modules archive/zip and Request.ParseForm, with identical updates also provided for Go 1.25.6 and Go 1.24.12. <a href="https://app.daily.dev/posts/DFYbtJ1rQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks and Snowflake face market maturation, pivot to AI</strong>: Databricks and Snowflake are reportedly facing growth saturation in their core data platform markets and are pivoting heavily into AI and applications to sustain high valuations, amidst increasing competition and customer demand for open data formats like Iceberg. <a href="https://app.daily.dev/feed-by-ids?id=Kv2NZR8R1&amp;id=LnOTpvngc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>OpenAI brings targeted ads to free ChatGPT users</strong>: OpenAI is now testing targeted ads in ChatGPT for free and Go tier users in the U.S., claiming these ads will be “relevant” and won’t influence chatbot responses. Because nothing says “unbiased AI” like a perfectly placed ad for that thing you just talked about. They assure us user data won’t be sold, which is comforting, coming from the company that just got sued by Elon Musk for allegedly abandoning its non-profit mission. <a href="https://app.daily.dev/posts/DQh9IAXNl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s AI search makes up crimes, Google says it’s ‘unavoidable’</strong>: Google’s AI search feature has been caught generating false accusations of criminal behavior against Norwegian politicians, causing severe reputational damage. Google’s defense? Errors are “unavoidable” given the scale of the internet. So, apparently, we should just accept that our AI overlords will occasionally invent felonies for us, and it’s totally fine because, you know, scale. <a href="https://app.daily.dev/feed-by-ids?id=507BvKbPR&amp;id=QiN4er8h3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wikimedia partners with AI firms, ignores bias concerns</strong>: Wikimedia Foundation is celebrating its 25th anniversary by adding six new AI companies, including Microsoft and Mistral AI, to its enterprise API program, giving them preferential access to Wikipedia content for training. This is great for revenue, but the persistent concerns about biased editing on Wikipedia impacting downstream AI systems? Apparently, that’s just a minor detail. Who needs accuracy when you have AI-powered training data? <a href="https://app.daily.dev/feed-by-ids?id=8kvN42B9c&amp;id=breqLrLKL" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
