October 19, 2026
The best of what first appeared in the trends on this day (UTC).
xiaoyang111222/cn-math-taxonomy
数学(小学到高中)、英语(三年级到高中)、初中物理的知识图谱:1644 个知识点、2403 条先修边、6723 道题,对齐课标,附自适应前置诊断。
A look at Anthropic's opaque, five-member Long-Term Benefit Trust, where members own no Anthropic equity but hold enormous sway by appointing most of its board
AI company has shared few details on the dynamics of the five-person trust empowered to shape its board and block major moves
Updates to site guidelines and AI/LLM tags
Hey folks, I've been having to perform an increasing number of [moderation actions](https://lobste.rs/moderations?moderator=355E3B) surrounding the `vibecoding` tag and AI/LLM related content. I started reading Lobsters because it had the most interesting viewpoints related to c…
Gradient Routing Localizes and Removes Unintended Behaviors in RL
Authors: Shawn Hu*, Jake Ward*, Aria Wong, Nathan Hu, Bryce Woodworth, Alexander Matt Turner, Alex Cloud (*Equal contribution, randomly ordered) tl;dr. During training, models sometimes learn to cheat at tasks rather than solve them in the…
Learning to Think and Explore in 3D World
ACT3D trains vision-language models via reinforcement learning to actively explore 3D environments from single first-person views, improving answer accuracy through adaptive, question-conditioned navigation without direct exploration…
Elon Musk says “I will give one of my SpaceX shares to every kid in America” in 2027, a gift worth ~$10.5B, after Anthropic committed 5% of its shares to APBF
Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.
Capital Ownership Will Make You Go Crazy First
People worried about a “permanent underclass” often assume that owning capital will protect them from disempowerment, or at least that they will be disempowered last, after the jobless masses dependent on UBI. I think something else is mor…
LiveMACE: Process-Aware Evaluation of LLM Agent Capabilities in Evolving Markets
LiveMACEBench evaluates persistent LLM agents in live financial markets using decision traces to reveal gaps between outcomes and underlying capabilities.
The US DHS orders all federal civilian agencies to remove DeepSeek's tech from their IT systems and prohibits future procurement of its products and services
The Department of Homeland Security has ordered federal civilian agencies to remove artificial-intelligence technology developed …
alchaincyf/huashu-motion
让你的agent一键直出商业级产品宣传片:会作曲,能指着画面改。Commercial-grade launch films by your coding agent — point at the frame to change it.
Linux 7.3
Linus Torvalds has released the 7.3 kernel, as expected. It's Sunday afternoon, and this wee [...]
Pack Light, CarryOn: Building Object-Oriented Context for Long-Horizon Agents
CarryOn organizes long-horizon agent context around reusable task objects rather than interaction histories, improving efficiency and performance on complex benchmarks.
Apple announces a “Pro and tell” event on October 28 at 10am PT in LA, where its first touchscreen MacBook, a 14-inch MacBook Pro with M6, and more are expected
Apple today announced its second experience of the month, taking place next Wednesday, October 28 at 10 a.m. PT. The tagline for this event is: “Pro and tell.”
Making all divisions slower to fix a bug in 0.000000000000000000542% of cases
A high-performance and high-level purely functional data-parallel array programming language that can execute on the GPU and CPU.
huggingface.wtf, an accessible but complete explanation
Announcing www.huggingface.wtf, a site that explains: "Hugging Face. Wtf?" I wrote the most accessible-but-complete explanation I could. Along the way, I made a nice AI Quotes database that I think will be useful for people who already und…
Timestral: Mistral's time-series foundation models
State-of-the-art probabilistic forecasting in four sizes (5M–209M params), CPU-efficient, zero-shot ready.
Has GPT-6-Astra Achieved Embodied Intelligence?
GPT-6-Astra achieves high performance across diverse robotic tasks via in-context learning, yet real-world deployment reveals gaps in demonstration cost, execution time, and practical embodied intelligence.
Sources: Nscale faces investor pushback over its ~$30B IPO valuation, with some urging a lower $16B-to-$19B valuation, as Microsoft exits a UK data center deal
Nscale Ltd. is facing opposition to the valuation target it's pursuing in its initial public offering, according to people familiar …
Twitch Plays: AI Interpretability of Bonsai-8B
Quick Links Watch the Twitch stream. Check the status of the interpretability map. Do your own parallel research and submit pull requests to fill out the map. What is this? Let's Play Interpretability is a Twitch livestream where you can h…
HexaShell: A Thin Harness for Exploring the Limits of VLM Agents as Robot Coding Agents
HexaShell is a coding agent that controls a dual-arm robot via vision and self-written tools, achieving state-of-the-art results on RoboDojo but struggling with contact-heavy tasks.
MiniMax launches M3.1 Flash, saying it delivers “a generational leap” in agentic tasks, coding, computer use, and cybersecurity, with weights due October 23
Space Bunny 🐰, which processed more than 200 trillion tokens while operating anonymously, was an early version of MiniMax M3.1 Flash.
Ansub/tv-detox
Make an old Android TV fast again with Claude Code or Codex: a tiny ad-free home screen plus the prompt to clean it up
Ask HN: How to deal with shipping AI-written code you don't understand?
I saw this LinkedIn post from a Meta engineer and it resonated with me: https://www.linkedin.com/posts/ali-hamie_im-an-engineer-at-m... > I'm an engineer at Meta and I never felt this uncomfortable doing my job. I'm writing code these days…
Scene-Designer: Learning Design Intelligence for Agentic 3D Scene Generation
Generating complete, coherent, and usable 3D indoor scenes from natural language requires strong design intelligence: the ability to design high-quality scenes and critique their quality. However, existing approaches either rely on…
Sources: DHS officials asked CISA to hack a voting machine to report vulnerabilities to the White House without telling the vendor; the request wasn't acted on
Reporting Highlights — An Unusual Request: Political appointees at the Department of Homeland Security wanted the agency …
Sorry, you are still dead
Epistemic status: A brief exploration based on court documents and contemporary reporting. My interpretation is tentative. AI use: context research, grammar. I don't remember how I stumbled upon the topic of legally dead people who are sti…
August in Servo: CSS Grid, text selection, and more
Plus accessibility and navigation improvements.
Starcraft: Ashfall
A fan-made StarCraft game. Completely free, on web and mobile. Command Terran and Protoss forces, hold the bridge, and see how deep the swarm really goes.
BizBench: Unveiling Hidden Challenges in Deploying Long-Horizon Agents for Real-World Business
BizBench evaluates long-horizon business agents on intermediate execution validity using real-world payment procedures, revealing that frontier models frequently violate constraints despite achieving partial goal success.
Filing: US prosecutors say OpenAI suspect Abdulrahman Alfozan maintained frequent WhatsApp contact in 2025 and 2026 with an unnamed Saudi Royal Family member
The Saudi national charged with planting a recording device in a sensitive OpenAI executive conference room in San Francisco was …
I no longer care to work on any Ruby gems so I've archived all my repositories
I’ve decided to archive the source code of my Ruby gems to make it clear that I’m no longer working on them; I’ve written this post to link from them in order to explain why.
I Can See Everything
Nick Cammarata of Goodfire makes a bold claim: Neural Networks are Already Clean. This was his follow up after various resulting online discussions. This is clearly good work that tells us new and important things. The question is, how big…
BridgeWAM: Bridging Video and Action Experts via Latent Bridge Queries
BridgeWAM decouples video and action experts via latent bridge queries to improve action prediction efficiency and generalization in world action models.
Microsoft debuts new security tools in Microsoft 365 E7 and its Project Perception system that treat AI agents as potential insider threats via risk assessments
Microsoft is coming out with new security tools that treat AI agents as potential insider threats, continuously assessing …
Tor VPN Beta: What we've learned building our own VPN for Android from scratch
Last fall, we soft-launched Tor VPN Beta as a way to extend Tor's privacy protections beyond the browser to an entire Android device. Since then, we've been able to learn from how people are actually using it in real-world scenarios. This…
humblemuzzu/dumbtv
Take the ads, tracking and clutter out of your Fire TV over ADB, no root. Every step can be undone. Comes with its own open-source home screen.
Lexical Reference Frames: Interpreting LM Residual Streams in Native Semantics
Lexical reference frames interpret hidden states through a model's own remaining layers, enabling accurate attribution, steering, and early layer exit.
Researcher: a Chinese-speaking hacker used AI agents to target 1,200+ websites and breach 100+ entities, like Deere and SharkNinja, spending ~$5,000 in tokens
Over the past couple of weeks, a cybersecurity researcher watched as a Chinese-speaking hacker broke into websites belonging …
Beginning of the Ende (a community Deno fork)
The trend of VC-backed tooling ending in abandonment rears its ugly head again. But hey, people gotta eat, right? The community will pick up the slack and steward the projects, right?
Android apps running native on macOS
Mac.apk: Run Android apps natively on Apple Silicon Macs - kksimp/Mac.apk
Into the Unknown: Benchmarking AI agents on open scientific research problems
Researchers benchmarked AI agents on open scientific problems by having them draft follow-up papers from unpublished directions, finding that none met expert standards and most contained fabricated claims or contradictory evidence.
Developers and vibe coders are using AI tools like Claude Code to remove sponsored ads on their smart TVs, speed up app launchers, disable tracking, and more
The latest use case for AI? Dumbing down your smart TV to get rid of the bloat. A number of developers and vibe coders …
Alternatives to animal testing are finally taking off
Changing habits may be the hardest engineering challenge of all
AI as a Normal Technology should be a vision
"AI as a Normal Technology" (AINT), proposed as a counterweight to "AI as an Abnormal Technology" (AIAbT) makes much more sense as a prescriptive / normative framework than a descriptive one. To be clear, the fact that they could be differ…
If you don't run your own LLM proxy, how do you know whats going on?
I think the title captures the idea fully but I'll dig into it because this ain't Twitter. You're working an idea, a piece of code, or a multibillion a year SaaS app which is probably saving the world. Maybe you're automating the cloning…
Xiaomi-Lens-0 Technical Report
Xiaomi-Lens-0 is a 4B multimodal model that predicts masks from internal visual representations to improve fine-grained region reasoning and referring comprehension.
Town, which develops enterprise AI assistants, raised a $108M Series B at a $1B valuation, after announcing a $55M Series A in June
Town, a two-year-old, San Francisco-based startup developing AI assistants, nabbed a unicorn valuation on Monday, announcing a $108 million Series B funding round co-led …
Superoptimizing an FPU emulator
Superoptimization is the process where a compiler automatically finds the optimal sequence for a loop-free sequence of instructions. – Wikipedia Compilers are good at making mechanical transformations that make code faster. Unfortunately…
William Gibson and the trademarking of cyberspace
Autodesk did trademark cyberspace, and William Gibson did get mad. But the story that Gibson retaliated by trademarking Eric Gullichsen was a prank by Timothy Leary.
Putting Linux on my RTX Spark laptop because I don't have a Microsoft account
Arch Linux ARM on the ASUS ProArt P16, with a bootable installer, published Windows comparisons, and local Qwen inference.
OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools
OpenMobile-2 is an open mobile agent trained with diverse simulated apps, large-scale supervised and reinforcement learning data, and hybrid GUI-tool use, achieving strong benchmark results and real-device transfer.