<h2>Must know</h2>
<ul>
<li><strong>NVIDIA and Lilly team up for AI drug discovery</strong>: NVIDIA and pharmaceutical giant Lilly announced a $1 billion AI co-innovation lab at the J.P. Morgan Healthcare Conference. The lab aims to revolutionize drug discovery by integrating computational modeling with experimental wet labs, leveraging Lilly’s biopharmaceutical expertise and NVIDIA’s AI infrastructure, including the DGX SuperPOD. This initiative seeks to model biological complexities more accurately and accelerate the discovery of new molecules, as reported by NVIDIA. <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 mathematical challenge that remained open until now. 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 rare clarity in attributing a new mathematical discovery primarily to artificial intelligence. <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’ by IDC MarketScape for its Unified AI Governance Platforms, integrating responsible AI standards across its ecosystem. The platform offers centralized oversight for the entire AI lifecycle, including generative and agentic AI, with features like real-time threat detection and automated compliance. Microsoft’s marketplace now boasts over 11,000 AI models and 4,000 AI applications, including pre-vetted solutions from Anthropic, OpenAI, Meta, and NVIDIA, aiming to accelerate enterprise AI adoption with a focus on compliance and security. <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 native GeForce NOW apps for Linux PCs (Ubuntu 24.04+) and Amazon Fire TV devices, enabling high-resolution, high-framerate cloud gaming powered by RTX 5080 GPUs. <a href="https://app.daily.dev/posts/5ceYZHV0v" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents in gaming environments show promise and limitations</strong>: Experiments with AI agents in Pokémon Showdown and RollerCoaster Tycoon 2 demonstrate AI’s ability for structured decision-making and park management, but highlight struggles with domain-specific and spatial reasoning. <a href="https://app.daily.dev/feed-by-ids?id=6IAZSWIz1&amp;id=NVzHSqkMx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA enhances enterprise AI security with BlueField</strong>: NVIDIA updated its Enterprise AI Factory design to include BlueField cybersecurity and acceleration, offering hardware-accelerated, zero-trust security for AI workloads with integrations from major security vendors. <a href="https://app.daily.dev/feed-by-ids?id=SCEwsKfyS&amp;id=WxRVVmFMH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ClickHouse acquires Langfuse for LLM tracing</strong>: ClickHouse acquired Langfuse, an open-source platform for LLM tracing and evaluation, aiming to enhance its development while maintaining open-source commitments and leveraging ClickHouse’s infrastructure for performance. <a href="https://app.daily.dev/feed-by-ids?id=O5KxwsHsv&amp;id=dliP781v1" 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 supporting inductive data types and pattern matching. <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 targets embedded systems</strong>: Porffor is a new ahead-of-time JavaScript compiler aiming for WebAssembly and native code, promising significantly smaller binaries and speed improvements for serverless, game engines, and embedded systems, though it currently lacks garbage collection and web API support. <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>Shopify deprecates tracking cookies, requires Hydrogen storefront updates</strong>: Shopify will deprecate shopify_y and shopify_s tracking cookies in 2026, requiring Hydrogen storefronts using React Router on Oxygen to upgrade by April 30, 2026, to maintain accurate visitor and session attribution for Shopify Analytics. <a href="https://app.daily.dev/feed-by-ids?id=QzarOUOVo&amp;id=yU3ZcbWbA&amp;id=zoNDLvNFD" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents are just APIs, nothing has changed</strong>: Apparently, ‘AI agents’ are the new hotness, but according to some technical folks, it’s just a fancy rebrand of good old API and web automation. The hype, they say, is mostly investor-driven fluff for financial gain, not some groundbreaking technological leap. So, while the VCs are busy chasing the next shiny object, the rest of us are still just calling functions, but now with more ‘agentic’ flair. <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>AI coding assistants: the ultimate productivity hack (or not)</strong>: One Google engineer claimed Claude Code did in an hour what took their team a year, only to later clarify it was a ‘simplified toy version.’ Meanwhile, other developers report AI-generated code leading to architectural nightmares, subtle bugs, and refactoring taking longer than writing from scratch. It seems AI is great for ‘productivity’ if you enjoy debugging hallucinations and rebuilding everything later. Who needs clean code when you have ‘velocity’? <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>The AI data center boom: building the future with no one to build it</strong>: The U.S. is apparently going to need 106 GW of power for AI data centers by 2035, and a casual 400,000 more construction workers by 2033, especially electricians and engineers. Companies are ‘recruiting’ from nuclear and military sectors, because apparently, building server farms is just like disarming a bomb, but with more concrete. Good luck with that ‘future of AI’ when there’s no one left to wire it up. <a href="https://app.daily.dev/posts/JE89HgTw1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
