<h2>Must know</h2>
<ul>
<li><strong>NVIDIA and Lilly team up for AI drug discovery</strong>: NVIDIA and pharmaceutical giant Lilly are investing $1 billion in an AI co-innovation lab to revolutionize drug discovery. Announced at the J.P. Morgan Healthcare Conference, the lab will combine computational modeling with experimental wet labs, leveraging NVIDIA’s AI infrastructure, including the DGX SuperPOD with B300 systems. This initiative aims to model biological complexities more accurately and accelerate the discovery of new molecules, as reported by NVIDIA and Lilly. <a href="https://app.daily.dev/posts/TcHQkCPbq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI solves a long-standing math problem</strong>: GPT 5.2 Pro, an AI model, has reportedly solved Erdos problem #281, a previously unsolved mathematical challenge. Mathematician Terence Tao, a highly credible source, stated this is “perhaps the most unambiguous instance” of AI solving an open mathematical problem. This marks a significant milestone for AI in mathematics, signaling a new era for AI-assisted research. <a href="https://app.daily.dev/posts/pRkd3CPEF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft pushes unified AI governance and marketplace</strong>: Microsoft has been recognized as a leader in Unified AI Governance Platforms by IDC MarketScape for 2025-2026, integrating responsible AI standards across its ecosystem. The platform offers centralized oversight for the entire AI lifecycle, including generative and agentic AI, with real-time threat detection and compliance for over 100 regulatory frameworks. Additionally, Microsoft’s Marketplace now features over 11,000 AI models and 4,000 AI applications, aiming to accelerate enterprise AI adoption with a focus on security and compliance. <a href="https://app.daily.dev/feed-by-ids?id=VuTiPUrsF&amp;id=XdBf9IqBR" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>NVIDIA expands GeForce NOW to Linux and Fire TV</strong>: NVIDIA announced at CES 2026 that GeForce NOW is expanding with native apps for Linux PCs (Ubuntu 24.04+) and Amazon Fire TV devices, offering up to 5K resolution at 120fps or 1080p at 360fps streaming, powered by GeForce RTX 5080 GPUs. <a href="https://app.daily.dev/posts/5ceYZHV0v" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA enhances enterprise AI security with BlueField</strong>: NVIDIA’s Enterprise AI Factory validated design now includes BlueField cybersecurity and acceleration features, offloading networking, storage, and security tasks to dedicated processors for hardware-accelerated, zero-trust security in AI workloads. <a href="https://app.daily.dev/posts/SCEwsKfyS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents in software development: mixed results</strong>: Developers are finding AI coding agents useful for structured tasks like blog migration, but unsupervised use often leads to technical debt, architectural errors, and misleading code, especially with complex applications. <a href="https://app.daily.dev/feed-by-ids?id=6XeVWOBXO&amp;id=STWE5tRXe&amp;id=UAphs9F5B&amp;id=euzGVQGzQ&amp;id=iukrPGTg6&amp;id=m6P4U0J0b&amp;id=wCQfiCXDa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA DRIVE powers Mercedes-Benz CLA and expands ecosystem</strong>: The new Mercedes-Benz CLA, set for late 2026 US release, will feature NVIDIA DRIVE AV software for advanced driver assistance, while NVIDIA also expanded its DRIVE Hyperion ecosystem at CES 2026 with new partners and the powerful DRIVE AGX Thor SoCs for level 4 autonomous driving. <a href="https://app.daily.dev/feed-by-ids?id=kCsBTg7xZ&amp;id=pTEo8FtLn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New programming language Kip uses Turkish grammar</strong>: Kip is an experimental programming language that uses Turkish grammatical cases in its type system to specify argument relationships in functions, allowing for flexible argument ordering and exploring the fusion of linguistic morphology with type theory. <a href="https://app.daily.dev/feed-by-ids?id=f4hbmP6tB&amp;id=kTJVodinv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Porffor JavaScript compiler and Xous microkernel OS emerge</strong>: Porffor is a new ahead-of-time JavaScript compiler targeting WebAssembly and native code for smaller, faster binaries, while Xous is a microkernel-based OS for embedded systems emphasizing strict process isolation and minimal kernel complexity. <a href="https://app.daily.dev/feed-by-ids?id=xTAvQVuAU&amp;id=xU3ZYAZlY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Curl project drops OpenSSL-QUIC backend</strong>: The curl project is removing support for the OpenSSL-QUIC backend in version 8.19.0 due to its inadequate API, extremely poor performance (up to 3x slower), and excessive memory usage (up to 20x more than 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 agents: just old automation with a new hat</strong>: Apparently, the tech industry has discovered ‘AI agents,’ which, according to some ‘experts’ and ‘historians,’ are just rebranded API/web automation. It seems the ‘agent hype’ is less about groundbreaking technology and more about convincing investors that we’ve finally cracked the code on something truly revolutionary. Because, you know, calling it ‘automation’ just doesn’t fetch those sweet, sweet VC dollars anymore. <a href="https://app.daily.dev/feed-by-ids?id=4rJ96gwiU&amp;id=J1BKrq2u6&amp;id=JE89HgTw1&amp;id=JItr1KihQ&amp;id=bXoqeH8eP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA’s latest gaming tech: more hype, less independent verification</strong>: NVIDIA just dropped a slew of announcements, including DLSS 4.5, G-SYNC Pulsar, and AI enhancements, all promising to ‘supercharge gameplay’ and deliver ‘1,000Hz+ motion clarity.’ Conveniently, all these claims, benchmarks, and performance metrics come directly from NVIDIA’s own marketing materials. It’s almost as if they want us to believe their hype without any pesky independent assessments or third-party reactions. Who needs objective data when you have hyperbolic language? <a href="https://app.daily.dev/posts/jXjzYY9vI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI in gaming: still struggling with basic spatial reasoning</strong>: Experiments with AI agents in games like Pokémon Showdown and RollerCoaster Tycoon 2 show that while they can handle rule-based reasoning, they consistently fall flat on tasks requiring domain-specific or spatial understanding. Claude Code, for instance, struggled with pathway layouts and rollercoaster construction. So, while AI can play Pokémon, don’t expect it to design your next theme park without turning it into a death trap. Maybe stick to the humans for anything that requires actual common sense. <a href="https://app.daily.dev/feed-by-ids?id=6IAZSWIz1&amp;id=NVzHSqkMx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
