<h2>Must know</h2>
<ul>
<li><strong>OpenAI goes all-in on ads with ChatGPT Go</strong>: OpenAI has launched ChatGPT Go, an $8/month subscription tier with ads, globally, following its India debut. This aggressive monetization strategy is a direct response to Google Gemini’s market share growth, aiming for $1 billion in ad revenue in 2026. Critics, including ‘The A in AGI stands for Ads,’ highlight privacy concerns as OpenAI plans to leverage user conversation data for ad targeting, despite assurances of privacy. This move signals a shift towards an ad-funded model for AGI development, mirroring Google’s historical strategy of low initial prices followed by gradual hikes. <a href="https://app.daily.dev/feed-by-ids?id=ArnlILqap&amp;id=ep7p1hnQy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>China’s massive satellite internet ambitions challenge Starlink</strong>: China has filed plans with the ITU to launch two mega-constellations, CTC-1 and CTC-2, totaling nearly 200,000 satellites, dwarfing Starlink’s projected 49,000. This move, if realized, could significantly alter global internet access and control, prompting the U.S. FCC to accelerate approvals for domestic satellite networks due to national security concerns. Meanwhile, Iran successfully disrupted Starlink services using military-grade GPS jamming, degrading performance by up to 80% and raising questions about the system’s resilience against state-level electronic warfare, as reported by multiple sources. <a href="https://app.daily.dev/feed-by-ids?id=9Dqo8zul1&amp;id=VVK9EFJ8w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI code generation still struggles with basic memory bugs</strong>: AI code generation tools consistently reproduce a memory management bug involving buffer allocation with <code>malloc</code>, returning the allocated buffer size instead of the actual decompressed data. This can lead to security vulnerabilities by exposing uninitialized memory, according to an analysis. The persistence of this flaw highlights a systemic limitation of LLMs, which tend to statistically mirror patterns—even erroneous ones—from their training data, suggesting that current statistical AI approaches may not easily eliminate such fundamental mistakes. <a href="https://app.daily.dev/feed-by-ids?id=mY0RAoEVz&amp;id=x9XQt7sEo" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Wine 11.0 stable release brings major architectural improvements</strong>: Wine 11.0 is out with 6,300 code changes, a new NTSYNC kernel module for better synchronization, a complete WoW64 architecture overhaul, and EGL as the default OpenGL backend, improving graphics and 32-bit app compatibility on 64-bit systems. <a href="https://app.daily.dev/posts/5LDn1X8et" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Figma control from the command line for AI agents</strong>: A new CLI tool, <code>figma-use</code>, offers full read/write access to Figma, allowing AI agents and developers to programmatically create designs using JSX syntax, optimized for token efficiency and offering over 100 commands. <a href="https://app.daily.dev/posts/DfAoRLrdU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Jakarta EE 12 development progresses with community events</strong>: The Jakarta EE Platform project is moving towards Milestone 2 for Jakarta EE 12, targeting January 27, 2026, with JakartaOne Livestream 2025 videos now public and several community conferences scheduled. <a href="https://app.daily.dev/posts/YrhkKGKoN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ReactOS gets massive networking performance boost</strong>: ReactOS has merged a long-awaited asynchronous TCP support patch, significantly improving networking performance for applications like browsers and FTP clients after 10 years of development. <a href="https://app.daily.dev/posts/wwS8HD8X1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Curl drops OpenSSL-QUIC backend due to poor performance</strong>: The curl project is removing support for the OpenSSL-QUIC backend in version 8.19.0, consolidating HTTP/3 support to nghttp2+nghttp3 and quiche due to OpenSSL-QUIC’s inadequate API, extremely poor performance, and excessive memory usage. <a href="https://app.daily.dev/posts/As2f3e0AW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pharmaverse blog speeds up CI/CD with Docker containers</strong>: The pharmaverse blog reduced its CI/CD publishing time from 17 minutes to 5 minutes by switching from installing R packages during each GitHub Actions run to using a pre-built Docker container image. <a href="https://app.daily.dev/posts/ZDzd1Eyhy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical RCE vulnerability in n8n automation platform</strong>: A critical unauthenticated remote code execution vulnerability (CVE-2026-21858) in n8n allows attackers to read arbitrary files and execute commands, affecting an estimated 100,000 servers globally; users must upgrade to version 1.121.0 or later. <a href="https://app.daily.dev/posts/4XP5JDLv1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The ‘Dead Internet Theory’ and AI’s authenticity crisis</strong>: Apparently, since 2016, the internet has been mostly bots and AI, according to the ‘Dead Internet Theory’ sparked by a HackerNews discussion. This theory, fueled by AI-generated code and comments, suggests we’re all just talking to machines. Because, you know, it’s not like humans ever produced questionable content or repetitive comments before AI came along. It’s definitely not just a convenient excuse for why online discourse feels a bit… off. <a href="https://app.daily.dev/feed-by-ids?id=JjrjV7tSv&amp;id=ZeB3UcRxG&amp;id=alpyX7cTa&amp;id=m1Q81Q3xQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-powered cat weather forecasts: because why not?</strong>: Someone built a DIY weather display that uses Google’s Gemini and Nano Banana APIs to generate AI-powered cat scenes based on the daily forecast. Because what we really needed was more AI-generated fluff, now with added feline charm, to tell us it’s raining. At 14 cents per generation after free credits, it’s a purr-fectly reasonable expense for such groundbreaking technology. <a href="https://app.daily.dev/posts/4SyrvZdf0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents for coding: the joy is gone, but at least it’s ‘good enough’</strong>: A developer, after 20 years of Neovim mastery, is now questioning the point of learning fundamental tools like <code>awk</code> and <code>sed</code> because AI coding agents are ‘good enough.’ The piece laments the loss of joy from the craft, implying that efficiency gains come at the cost of satisfaction. So, we’re trading the joy of creation for the lukewarm comfort of ‘good enough’ AI. What a time to be alive. <a href="https://app.daily.dev/posts/vFzyyNK3z" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
