<h2>Must know</h2>
<ul>
<li><strong>AI solves a previously unsolved math problem</strong>: GPT-5.2 Pro has reportedly solved Erdos problem #281, a mathematical challenge that human mathematicians had not cracked. Terence Tao, a highly respected mathematician, called this “perhaps the most unambiguous instance” of AI solving an open problem, marking a significant milestone for AI in mathematical research. This achievement, reported on X, highlights the growing capabilities of AI in complex problem-solving. <a href="https://app.daily.dev/posts/pRkd3CPEF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>China plans a massive satellite constellation</strong>: China has filed requests with the ITU for two new satellite constellations, CTC-1 and CTC-2, totaling an astounding 193,428 satellites—four times larger than SpaceX’s Starlink plans. This move, filed by a newly established Chinese entity, has raised national security concerns for the US FCC, which is accelerating approvals for American satellite networks in response. While China’s current deployment is limited, the sheer scale of this proposal suggests a significant strategic play, as reported by various sources. <a href="https://app.daily.dev/posts/9Dqo8zul1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Snap Store domain hijacking for malware distribution</strong>: Scammers are exploiting the Canonical Snap Store by hijacking expired domains of legitimate publishers to distribute malicious updates to trusted snap applications. This new tactic, an escalation from previous methods, specifically targets cryptocurrency wallet users to steal recovery phrases. The attack undermines trust in Snap publisher reputation signals and highlights critical vulnerabilities in authentication tied to domain registrations, as detailed in cybersecurity reports. <a href="https://app.daily.dev/posts/8RJLmL5a0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Physical AI and AV industry shifts at CES 2026</strong>: CES 2026 saw Physical AI as a central theme, with Mobileye acquiring Mentee Robotics for $900 million, Motional shifting to an “AI-first” AV development, and Tesla moving to a subscription-only model for Full Self-Driving, while Chinese automakers gained visibility amidst US regulatory tensions. <a href="https://app.daily.dev/posts/MtQzzuuIN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Iran disrupts Starlink internet with GPS jamming</strong>: Iran used military-grade GPS jammers to disrupt Starlink internet, cutting performance by up to 80% during a nationwide internet shutdown, raising concerns about Starlink’s reliability in conflict zones and its perceived censorship resistance. <a href="https://app.daily.dev/posts/VVK9EFJ8w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New CLI tool integrates AI and developers with Figma</strong>: The <code>dannote/figma-use</code> command-line tool offers full read/write access to Figma, enabling AI agents and developers to programmatically create and modify designs using JSX syntax, with an experimental ‘render’ command promising up to 100x faster operations. <a href="https://app.daily.dev/posts/DfAoRLrdU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open source models: cathedral, megachurch, and bazaar</strong>: Open source projects are categorized into three models: exclusive ‘cathedrals,’ large ‘megachurches’ like the Linux Foundation, and the numerous ‘bazaar’ projects, which form critical infrastructure but struggle with sustainable funding, prompting initiatives like the Sovereign Tech Agency’s funding models. <a href="https://app.daily.dev/posts/rr8Vo8FaG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s Antigravity expands to compete with Anthropic’s Co-work</strong>: Google’s Antigravity is now competing with Anthropic’s Co-work by integrating productivity features like presentation creation and file organization directly into the IDE, leveraging Gemini 3 Flash for non-coding tasks and aiming to keep developers within their editor for all work. <a href="https://app.daily.dev/posts/HVlI4ECqg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vertex AI and Claude pricing models revealed</strong>: Vertex AI and Claude have detailed their 2026 pricing, with Vertex AI offering various Gemini models priced per million tokens and custom training per node-hour, while Claude provides tiered subscriptions and token-based API pricing, both emphasizing cost optimization strategies. <a href="https://app.daily.dev/feed-by-ids?id=AslvqVnql&amp;id=77ufsy3OX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Curl drops OpenSSL-QUIC backend for HTTP/3</strong>: The curl project is removing support for the OpenSSL-QUIC backend in version 8.19.0 due to its inadequate API, poor performance (3x slower), and excessive memory usage (20x more) compared to alternatives, consolidating HTTP/3 support to nghttp2+nghttp3 and quiche. <a href="https://app.daily.dev/posts/As2f3e0AW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-driven security: more tools, same problems?</strong>: Apparently, 2025 saw a record 48,000 CVEs, with attackers exploiting vulnerabilities within 24 hours, thanks to AI. Now, we’re told the solution for 2026 is more AI-driven monitoring and memory-safe languages like Rust, which supposedly have 1,000x fewer vulnerabilities than C/C++. So, we’re fighting AI with AI, and somehow, WordPress plugins are still the biggest problem. Sounds like a perpetual motion machine for cybersecurity vendors. <a href="https://app.daily.dev/feed-by-ids?id=Tk2pS5qPG&amp;id=1GWtcQWuJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The GitHub contribution graph: a meaningless metric</strong>: In a shocking revelation, it turns out your GitHub contribution graph, LinkedIn profile, and LeetCode scores are ‘poor indicators of developer worth.’ Who knew that standardized metrics might not capture the nuanced reality of human value? The author suggests that true assessment requires ‘trust and understanding individual circumstances,’ which, let’s be honest, sounds a lot like ‘we don’t want to pay you what you’re worth if you don’t fit our narrow, algorithmic mold.’ <a href="https://app.daily.dev/posts/WQGtqJsDu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding agents: the joy killers?</strong>: After 20 years of Vim/Neovim, a developer is now questioning if learning fundamental tools like awk and sed still makes sense because AI coding agents are ‘good enough.’ The piece laments the loss of joy from the craft of coding, wondering what’s left when automation takes over. It’s almost as if the ‘efficiency gains’ promised by AI come with the hidden cost of turning creative problem-solving into glorified prompt engineering. Who needs joy when you have productivity metrics, right? <a href="https://app.daily.dev/posts/vFzyyNK3z" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
