<h2>Must know</h2>
<ul>
<li><strong>TSMC sees endless AI chip demand, plans massive expansion</strong>: TSMC reported record Q4 earnings for 2025 and forecasts 30% revenue growth for 2026, driven by insatiable AI chip demand. CEO C.C. Wei called AI a “megatrend” with “endless” chip demand, according to TSMC’s official financial reporting. The company plans $52-56 billion in capital expenditures for 2026, focusing 70-80% on advanced 7nm and smaller nodes, including a rapid ramp-up of 2nm production with 30% based in the USA. Major customers like Apple, Nvidia, and AMD are booking production slots years in advance, signaling sustained confidence in the AI sector. <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. This flaw, with a CVSS score of 10.0, stems from 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>AI infrastructure complexity and technical debt hinder enterprise ROI</strong>: Over half of US enterprises report AI project delays or cancellations due to infrastructure issues, with 66% finding their AI environments excessively complex, according to a DDN-commissioned survey. MIT research further indicates 95% of organizations see zero measurable returns from generative AI. Companies with unaddressed technical debt are far less likely to benefit, while those with modernized applications and integrated security are three times more likely to achieve clear ROI. The focus is shifting from simple AI adoption to deeper operational integration, emphasizing workforce education and smarter use cases over cloud migration. <a href="https://app.daily.dev/feed-by-ids?id=k12AAd1Ej&amp;id=tr8ZpGHVk" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>GitHub Copilot gets agentic memory</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>OpenAI to test targeted ads in ChatGPT</strong>: OpenAI will begin testing targeted ads in ChatGPT for free and Go tier users in the U.S., with ads appearing at the bottom of conversations and users retaining the ability to dismiss or disable personalization. <a href="https://app.daily.dev/posts/DQh9IAXNl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google BigQuery AI translates comments to SQL</strong>: Google is previewing a new BigQuery feature, “Comments to SQL,” powered by Gemini, which translates natural-language instructions within SQL comment blocks into executable queries 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, launches native Postgres service</strong>: ClickHouse acquired Langfuse, an open-source LLM observability platform, to enhance its AI application monitoring capabilities and launched a native Postgres service in partnership with Ubicloud for unified transactional and analytical workloads. <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>Agentic AI for non-human identity and secrets management</strong>: Agentic AI systems are emerging to autonomously manage and secure non-human identities (NHIs) like applications and services, reducing breach risks and improving compliance through anomaly detection and continuous monitoring. <a href="https://app.daily.dev/feed-by-ids?id=NCcbuPG8X&amp;id=UAGIEGe4I" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Palo Alto Networks patches denial-of-service flaw</strong>: Palo Alto Networks released patches for CVE-2026-0227, a high-severity denial-of-service vulnerability in PAN-OS firewalls with GlobalProtect enabled, which could allow unauthenticated attackers to force devices into maintenance mode. <a href="https://app.daily.dev/posts/MtreEtTLc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks and Snowflake face market maturation</strong>: Databricks and Snowflake are confronting growth saturation in their core data platform markets, prompting both companies to expand into AI and applications to sustain high valuations amidst intensifying 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>Elon Musk’s OpenAI lawsuit goes to jury trial</strong>: Apparently, Elon Musk’s lawsuit against OpenAI and Microsoft, alleging a breach of OpenAI’s original nonprofit mission, is actually going to a jury trial. A federal judge rejected motions to dismiss, so now a jury gets to decide if OpenAI prioritized profits over its founding principles and if Microsoft was in on it. Because nothing says ‘cutting-edge AI’ like a courtroom drama over broken promises from a decade ago. <a href="https://app.daily.dev/posts/pwpD2QRjm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-powered coding tools: the new ‘vibe coding’ for disposable projects</strong>: Linus Torvalds apparently used “vibe coding” with AI for a hobby project, which is now being touted as a trend for “micro apps.” While some developers claim AI tools cut development time from a day to an hour, they still can’t handle basic resume functionality or robust error handling. So, it’s great for quick, disposable projects, but for anything serious, you still need a human to clean up the AI’s mess. Progress! <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>AI agents: just while loops with extra steps</strong>: Some ‘innovative’ LLM agent systems are being exposed as nothing more than glorified while loops running unit tests. One expert sarcastically points out that these ‘groundbreaking’ AI agents are essentially basic programming constructs that have existed for years. So much for the hype about sentient code; turns out it’s just a loop trying to pass its own tests. <a href="https://app.daily.dev/posts/2dSW1OELp" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
