<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, and forecasts 30% growth for 2026, driven by insatiable AI chip demand. CEO C.C. Wei called AI a ‘megatrend’ and expects demand to be ‘endless,’ according to TSMC’s official financial reporting. The company plans $52-56 billion in capital expenditures for 2026, with most of it going to advanced 7nm and smaller process nodes, including a rapid ramp-up of 2nm production. This indicates a strong, long-term commitment to cutting-edge manufacturing for AI infrastructure. <a href="https://app.daily.dev/posts/aTc9ls9yL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI to introduce targeted ads in ChatGPT for free users</strong>: OpenAI is starting to test targeted ads in ChatGPT for free and Go tier users in the U.S., beginning January 16, 2026. These ads will appear at the bottom of conversations, tailored to the topics discussed, as reported by multiple sources. OpenAI states that ads will be clearly labeled, won’t influence chatbot responses, and user data won’t be sold. This move aims to support free access to ChatGPT and generate revenue, with a new $8/month Go tier also expanding to the U.S. <a href="https://app.daily.dev/posts/DQh9IAXNl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Elon Musk’s lawsuit against OpenAI and Microsoft heads to jury trial</strong>: Elon Musk’s lawsuit against OpenAI and Microsoft, alleging a breach of OpenAI’s original nonprofit mission, is proceeding to a jury trial. A federal judge rejected motions to dismiss from both OpenAI and Microsoft, setting the stage for a trial in late April. The core of the dispute revolves around OpenAI’s shift to a for-profit model and Microsoft’s alleged role in that transition, with jurors tasked to determine if original promises were broken. <a href="https://app.daily.dev/posts/pwpD2QRjm" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI coding tools: productivity gains vs. quality concerns</strong>: While AI coding tools like Claude can significantly reduce development time for tasks like building a Python CLI, they often fall short on nuanced requirements like robust error handling or resume functions, requiring manual intervention. <a href="https://app.daily.dev/feed-by-ids?id=SqdiiasJQ&amp;id=YfWBhNQB7&amp;id=H6Mp947gD&amp;id=h6ZJabZ0D&amp;id=laMxMdAYV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub Copilot introduces agentic memory for paid plans</strong>: GitHub Copilot now offers an agentic memory feature in public preview for paid plans, allowing it to learn and retain repository-specific insights to improve 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>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. <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>Google previews BigQuery AI for natural language to SQL</strong>: Google is testing a ‘Comments to SQL’ feature in BigQuery, powered by Gemini, which translates natural-language instructions within SQL comment blocks into executable SQL queries to accelerate data analysis. <a href="https://app.daily.dev/posts/5AcBUpc2i" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Flutter refactors for modularity and leaner apps</strong>: Flutter is undergoing a core SDK refactoring to decouple Material and Cupertino design systems, aiming for independent versioning, improved tree-shaking, and a more universal rendering layer. <a href="https://app.daily.dev/posts/T07WzSMOC" 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 Modular DS WordPress plugin versions 2.5.1 and below allows unauthenticated attackers to gain instant admin access, with a patch (v2.5.2) now available. <a href="https://app.daily.dev/posts/D6wKV3WK1" 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>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-powered coding tools: the emperor’s new code?</strong>: Apparently, ‘vibe coding’ with AI is all the rage for ‘quick, disposable projects,’ but for anything serious, it’s still a hot mess. Research (unspecified, naturally) suggests developers are 19% slower with AI tools, and open-source maintainers are drowning in ‘low-quality submissions’ from AI-assisted contributions. So, while AI might help you whip up a micro-app for choosing restaurants, don’t expect it to write your next production-ready, bug-free masterpiece. <a href="https://app.daily.dev/feed-by-ids?id=H6Mp947gD&amp;id=h6ZJabZ0D&amp;id=laMxMdAYV&amp;id=S37WCeeOE&amp;id=nLHgst6Y1&amp;id=dc9HNh5yX&amp;id=9IenhmVcQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Enterprise AI: more hype than ROI?</strong>: A survey by DDN (who definitely isn’t trying to sell you infrastructure) claims over half of enterprise AI projects are delayed or canceled due to ‘complexity’ and ‘underutilized GPUs.’ Meanwhile, MIT (also not trying to sell you anything) found 95% of organizations see ‘zero measurable returns’ from generative AI. Turns out, throwing AI at your technical debt just makes your debt more ‘intelligent’ but no less burdensome. Who knew? <a href="https://app.daily.dev/feed-by-ids?id=k12AAd1Ej&amp;id=tr8ZpGHVk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wikimedia’s AI partnerships: training on volunteer labor, ignoring bias?</strong>: Wikimedia just added six more AI companies, including Microsoft and Mistral AI, to its enterprise API program, giving them ‘preferential access’ to Wikipedia content for training their commercial AI systems. So, while volunteer editors are busy curating knowledge, AI firms are hoovering it up to make a buck. And those ‘persistent concerns’ about biased editing on Wikipedia? Well, they might just become ‘persistent concerns’ in your AI-generated responses too. But hey, revenue! <a href="https://app.daily.dev/feed-by-ids?id=8kvN42B9c&amp;id=breqLrLKL" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
