<h2>Must know</h2>
<ul>
<li><strong>TSMC sees endless AI chip demand, plans massive investment</strong>: TSMC reported record Q4 earnings for 2025, with revenue up 25.5% year-over-year, driven by what CEO C.C. Wei calls an “endless” demand for AI-related chips. The company forecasts 30% revenue growth for 2026 and plans $52-56 billion in capital expenditures, primarily for advanced 7nm and smaller process nodes. TSMC expects its 2nm manufacturing process to ramp up faster than previous nodes, with about 30% of that production slated for the USA, as reported by TSMC’s official financial statements. <a href="https://app.daily.dev/posts/aTc9ls9yL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical WordPress plugin vulnerability grants admin access</strong>: A critical privilege escalation vulnerability (CVE-2026-23550) in the Modular DS WordPress plugin, affecting versions 2.5.1 and below, allows unauthenticated attackers to gain instant admin access. The flaw, with a CVSS score of 10.0, is due to faulty API routing logic. Active exploitation has been detected, and over 40,000 WordPress sites are estimated to be vulnerable. 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 patching. Prisma Access cloud customers are mostly automatically patched, but NGFW users must apply updates manually, according to Palo Alto Networks. <a href="https://app.daily.dev/posts/MtreEtTLc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Quarkus newsletter highlights Java 25 FFM API and event-sourced systems</strong>: The January Quarkus newsletter features guides on common developer questions, a tutorial on the unified Hibernate with Panache extension, and an exploration of Java 25’s Foreign Function &amp; Memory API for native image processing, alongside community discussions and events. <a href="https://app.daily.dev/posts/va4C2SHIt" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Spring Data releases maintenance updates</strong>: Spring Data versions 2025.1.2 and 2025.0.8 have been released, providing maintenance updates, bug fixes, and minor improvements for Java applications. <a href="https://app.daily.dev/posts/6KncWQqKv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub Copilot introduces agentic memory in public preview</strong>: GitHub Copilot now features agentic memory in public preview for paid plans, allowing it 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>Google previews BigQuery AI for natural language to SQL</strong>: Google is testing a new ‘Comments to SQL’ feature in BigQuery, powered by Gemini, which translates natural language instructions within SQL comment blocks into executable queries, aiming to accelerate data analysis. <a href="https://app.daily.dev/posts/5AcBUpc2i" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ClickHouse acquires Langfuse, expands AI observability</strong>: ClickHouse 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 remains open-source and a standalone cloud service. <a href="https://app.daily.dev/feed-by-ids?id=7DDF3YQda&amp;id=BqVJ6XPkS&amp;id=kU2RvmlCw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New releases in open source database tools</strong>: Portabase, a new self-hosted backup and restore tool for PostgreSQL, MySQL, and MariaDB, has been released with Rust-refactored agents, alongside credcheck v4.4, a PostgreSQL extension update fixing a backend crash. <a href="https://app.daily.dev/feed-by-ids?id=5HNbhZxB9&amp;id=7mmyYg2vc" 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 newer CPUs not affected by TAA and related vulnerabilities, such as Cooper Lake and Ice Lake, aiming to improve performance where safe. <a href="https://app.daily.dev/posts/l4qet2HfG" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>OpenAI’s ad strategy: ‘mission alignment’ or just more cash?</strong>: OpenAI is rolling out targeted ads for free and Go ChatGPT users in the U.S., claiming it’s to ‘support continued free access’ and ‘expand affordable AI access globally.’ They assure us ads won’t influence responses and user data won’t be sold, all while expanding their Go tier to $8/month. It’s almost as if ‘mission alignment’ now means ‘monetize everything, but promise it’s for a good cause.’ <a href="https://app.daily.dev/posts/DQh9IAXNl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-powered coding tools: the ‘vibe coding’ revolution that isn’t</strong>: Linus Torvalds apparently used ‘vibe coding’ with AI for a hobby project, and now some are hailing it as the future. Yet, research suggests developers are 19% slower with AI tools, and open-source maintainers are drowning in low-quality, AI-generated submissions. So, while AI might help you consolidate your SmugMug photos in an hour, don’t expect it to write your next production-ready microservice without a lot of manual cleanup. <a href="https://app.daily.dev/feed-by-ids?id=H6Mp947gD&amp;id=h6ZJabZ0D&amp;id=laMxMdAYV&amp;id=SqdiiasJQ&amp;id=YfWBhNQB7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s AI search: ‘unavoidable’ errors and zero accountability</strong>: Google’s AI search has been caught generating false and defamatory information about Norwegian politicians, including criminal accusations. Google’s defense? Errors are ‘unavoidable’ given the scale of the internet. Apparently, the company that organizes the world’s information can’t be bothered to ensure its AI doesn’t ruin lives, preferring to shift the burden of ‘critical evaluation’ onto users. Classic tech giant move: all the innovation, none of the responsibility. <a href="https://app.daily.dev/feed-by-ids?id=507BvKbPR&amp;id=QiN4er8h3" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
