<h2>Must know</h2>
<ul>
<li><strong>AI coding agents demand more upfront planning and energy</strong>: AI coding assistants, while seemingly fast, are creating a productivity paradox, requiring significantly more upfront architectural planning and increasing code review times by 91% in high-utilization teams. A 2025 study found developers using AI tools took 19% longer to complete tasks, as AI-generated code often contains security vulnerabilities (45%) and lacks contextual memory, leading to inefficient ‘zig-zag’ development. Furthermore, these tools are energy hogs; a median Claude Code session consumes 41 watt-hours, 138 times more than a typical chatbot query, with daily usage comparable to running a dishwasher, as reported by Tech Lead Digest. <a href="https://app.daily.dev/feed-by-ids?id=pYjmmMIkn&amp;id=ex1Cm3GIg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Creative organizations ban generative AI content</strong>: Major creative organizations are taking a firm stance against generative AI. The Science Fiction and Fantasy Writers Association (SFWA) has revised its Nebula Awards rules to disqualify any work created wholly or partially by large language models, following member backlash against a more lenient policy. Similarly, San Diego Comic-Con has updated its art show policy to ban all AI-generated material. These moves reflect a growing pushback from creators aiming to preserve the integrity and originality of human-authored content, despite ongoing debates about enforcement and the integration of LLMs into common tools. <a href="https://app.daily.dev/posts/32kw1Fo3F" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel gets continuity plan for Linus Torvalds’ repo</strong>: The Linux 6.19 kernel now includes an official continuity document outlining procedures if Linus Torvalds’ mainline Git repository becomes unavailable. Developed at the 2025 Linux Maintainer Summit, the plan specifies a 72-hour window for the Maintainer Summit organizer or the Linux Foundation Technical Advisory Board Chair to initiate succession. This proactive measure ensures long-term project continuity for the Linux kernel, with a two-week deadline for a transition plan to be communicated to the community. <a href="https://app.daily.dev/posts/EogQGzxtg" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>LLM inference hardware bottlenecks identified</strong>: LLM inference is bottlenecked by memory capacity and bandwidth during the autoregressive Decode phase, not compute, driving research into High Bandwidth Flash, Processing-Near-Memory, 3D memory-logic stacking, and low-latency interconnects for next-gen datacenter AI accelerators. <a href="https://app.daily.dev/posts/t0LNVAkom" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Turbopuffer ANN v3 achieves low-latency vector search</strong>: Turbopuffer’s ANN v3 now delivers 200ms p99 query latency at 1,000 queries per second over 100 billion vectors (200TiB), using hierarchical clustering, RaBitQ binary quantization for 16-32x compression, and selective reranking to maintain 92% recall. <a href="https://app.daily.dev/posts/EHAeemnne" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>MIT achieves lower error rate in superconducting memory</strong>: MIT researchers have developed a 4x4 superconducting nanowire memory array operating at 1.3 Kelvin, achieving a bit error rate of 10⁻⁵ and a functional density of 2.6 megabits per square centimeter, addressing major scalability and error challenges for quantum and low-energy computing. <a href="https://app.daily.dev/posts/UFLhnkgMo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Measurement-free fault-tolerant quantum computing demonstrated</strong>: Researchers have demonstrated a measurement-free approach to fault-tolerant quantum computing using trapped ions, bypassing slow and error-prone mid-circuit measurements to achieve universal logical operations and execute Grover’s algorithm on three logical qubits. <a href="https://app.daily.dev/posts/b76Xs6rDw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.0 gains display support for Snapdragon 8 Elite Gen 5</strong>: The MSM DRM kernel driver for Linux 7.0 is updated to include display support for Qualcomm Snapdragon 8 Elite Gen 5 chipsets and the older Adreno 225 GPU, with changes submitted by Rob Clark for next month’s merge window. <a href="https://app.daily.dev/posts/LVxtNaVxq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AMD submits graphics driver fixes for Linux 7.0</strong>: AMD has submitted graphics driver fixes for AMDGPU and AMDKFD drivers ahead of the Linux 7.0 kernel merge window, addressing issues across RDNA4 and older architectures, including user-queue handling, HDMI support, and DCN improvements. <a href="https://app.daily.dev/posts/cxhDtp9rh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>FOSDEM 2026 to focus on AI and open source sustainability</strong>: FOSDEM 2026, expecting over 10,000 attendees, will focus on the impact of AI on open source, funding sustainability, digital sovereignty, and contributor burnout, alongside celebrating KDE’s 30th anniversary and showcasing a new NVIDIA driver in Rust. <a href="https://app.daily.dev/posts/TL2tOfQ6b" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Agentic AI will magically solve all your compliance woes</strong>: Apparently, ‘Agentic AI’ is the new magic bullet for enterprise compliance and ESG reporting, promising to transform manual, fragmented processes into automated, unified systems. According to a flurry of identical articles, it’s all about minimizing risk and turning compliance into a ‘strategic business advantage.’ Conveniently, there’s no mention of who’s actually using this, or any pesky downsides, just a lot of talk about ‘workflow-native intelligence’ and ‘strengthening governance standards.’ Sounds like another solution looking for a problem, or at least a budget. <a href="https://app.daily.dev/feed-by-ids?id=Skhkgwj1Z&amp;id=qF5kIahRI&amp;id=uTj3yEY4b" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Encrypted quantum cloning protocol defies physics, apparently</strong>: A new quantum protocol is reportedly enabling ‘encrypted cloning’ of quantum states, supposedly circumventing the no-cloning theorem—a foundational principle in quantum information theory. This ‘breakthrough’ promises advancements in quantum data storage and security, yet the article conveniently omits any technical details, names of researchers, or experimental evidence. It’s almost as if they’re claiming to have invented a perpetual motion machine without showing the blueprints. Color me skeptical. <a href="https://app.daily.dev/posts/z7OYuDS7t" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ChatGPT and Claude are now citing ‘Grokipedia’</strong>: In a stunning display of AI’s critical thinking, both ChatGPT’s GPT-5.2 and Anthropic’s Claude are reportedly citing ‘Grokipedia’—Elon Musk’s ‘conservative-leaning’ AI-generated encyclopedia—as a source for obscure topics. This ‘encyclopedia’ has faced criticism for being inaccurate and biased, but hey, why let facts get in the way of a good AI-generated answer? It just goes to show, garbage in, garbage out, even when the garbage is ‘AI-generated.’ <a href="https://app.daily.dev/posts/SDpas3mop" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
