October 28, 2026
The best of what first appeared in the trends on this day (UTC).
Apple debuts a touchscreen MacBook Pro with a PROLED XDR tandem OLED display, light-up logo, and M5 Pro or M5 Max; the 14" starts at $2,899, the 16" at $3,399
The world's best pro laptop now features an industry-leading tandem OLED display, an intuitive Touch experience …
This is not sustainable
Cyber Defense teams are suffering a DDoS; vulnerability management is no longer able to keep up with the onslaught of findings, reports, nor hype. The current situation is simply not sustainable.
anhao/ai-pose-gallery-zh
中文版 AI 姿势参考图库 · 1439 组 / 2878 张图片 / 完整汉化 · 零依赖静态画廊
You can now "@Grok" on LessWrong
Or if you prefer, "@Claude", "@Opus", "@Astra" @Grok is a popular Twitter/X feature where if you tag Grok in a tweet, Grok will respond in a comment. The stereotypical use case is "@Grok is this true?" But you can also ask things like "How…
Apple's git still trusts the compromised eMudhra certificates
Eleven days after Apple was notified of the eMudhra CA compromise, the git, curl and python3 that Apple ships with macOS still trust the compromised certificates.
Google unveils Gemini 4 Carbon, its “strongest model with state-of-the-art cybersecurity” capabilities, ranked #1 on Artificial Analysis' new Intelligence Index
Gemini 4 Carbon is our strongest model with state-of-the-art cybersecurity and long horizon capabilities.
The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal Agents
The Trinity framework recursively improves multimodal agents by organizing tasks into a capability tree, generating simulated training worlds, and routing specialized harness modules to boost performance on diverse benchmarks.
Reconceptualize things as your top affordances
Everyone has some actions that come naturally to them, and some actions that are more unnatural, and some actions that wouldn’t even occur to them. For me, blogging comes naturally. Which is not to say that it’s easy, but the challenge of…
Evidence of a spying operation by ICE against left-wing social movements
ICE’s Homeland Security Investigations is engaged in far-reaching spying against left-wing activists that rivals the FBI and CIA crackdowns on the Civil Rights and anti-war movements.
Amazon unveils the $500 Echo Show 16, the $200 Echo Time alarm, and the $200+ Alexa Photo Frame, raises smart speaker prices, and previews Alexa Home Map
With its latest devices, including a smart bedside alarm clock, Amazon looks to cement Alexa's presence in every room of the home
Multiscale Policy Optimization for Long-Horizon LLM Agents
Multiscale Policy Optimization improves long-horizon agent training by aligning credit assignment and policy updates at the turn scale.
Streetlight, Torchlight and Nightvision
Note: We at PrincInt are launching a program teaching “Discernment” or “Taste” and generally spend a lot of our time in workshops and otherwise trying to teach academics how to think about AI such that their research ends up being useful.…
SK Hynix reports Q3 revenue up 328% YoY to ~$78.8B, above ~$74.07B est., and operating profit up 632% to ~$62.76B, above ~$57.78B est. and a company record
Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.
VibeVideoBench: Evaluating Multimodal Agents on Code-Driven Video Generation
VibeVideoBench evaluates multimodal agents that generate videos via code, finding that top systems meet core requirements on most tasks but frequently fail due to poor time management.
interplains/Zero-Breach-Toolkit
Free, open-source, offline checks for traces of known mercenary spyware. Works on Windows, macOS and Linux, Android phones and iPhone backups.
MacBook Pro with PROLED XDR Display and Touch
https://www.apple.com/newsroom/2026/10/apple-unveils-groundb...
Microsoft reports Q1 revenue up 19% YoY to $92.28B, above $90.69B est., Azure revenue up 48% to $33.17B, above $32.32B est., and net income up 32% to $36.63B
REDMOND, Wash. — October 28, 2026 — Microsoft Corp. today announced the following results for the quarter ended September 30 …
PDRVLA: A Vision-Language-Action Model with Physical Dynamics Representations
PDRVLA improves robot action generation by integrating a lightweight physical dynamics module that predicts future visual features with world-model supervision, avoiding online video generation latency.
Some thoughts on GRAFT and follow-up work
I'm speaking as myself here and not on behalf of any of my co-authors. A bunch of the stuff in this post is explicitly speculative. I've written this post as a way of centralizing some points I want to make in discussion/in response to oth…
Start Doing Graph Theory
An introductory textbook about using graph theory to solve puzzles, prove theorems, and everything in between.
AMD says it will invest $1B in Meta's equity offering, its first Helios rackscale systems are installed at Meta, and its EPYC CPUs power Meta's Muse AI agent
Nvidia and Advanced Micro Devices have said they will invest a total of $4 billion in Meta after the social media …
ChronoReplay: One Answer Interface Is Not Enough for Streaming Spatial Intelligence
ChronoReplay mitigates interface lock-in in spatial vision-language models by replaying heterogeneous spatial supervision chronologically, improving cross-task transfer and enabling efficient inference-time evidence selection.
The Grid, a factory-scale robot in Massachusetts that builds PCBs on demand
A factory-scale robot in Massachusetts that builds circuit boards on demand, in half the time of a typical Shenzhen quick-turn shop.
Lawsuit: the parents of a 31-year-old say OpenAI's GPT-4o guided her through an aggressive tapering plan off epilepsy meds, leading to her death in February
The chatbot helped a woman taper off a lifesaving drug and hide the plan from her doctors. — 1.3k
Hy-Nomas: Towards Free-form Interactive World Generation
Hy-Nomas is a video world model that uses causal attention and structured captions to follow multi-turn text and camera instructions while preserving interaction history.
Anthropic Files Its S-1
The full text is finally here, so we are giving it the full RTFB of model card treatment. First, a recap of the state of Anthropic finances, then a close reading of the relevant parts of the S-1. A lot of it is technical and uninteresting,…
Emacs Turns 50
How the MIT AI Lab’s text editor gave birth to the free software movement.
Artificial Analysis unveils the Intelligence Index v5 with new coding and science benchmarks; 60% of the Index is based on private test sets or held-out answers
The Artificial Analysis Intelligence Index is our composite measure of model capabilities across tasks reflecting real-world usage today. v5 …
Memory Tax: Dissecting the Net Value of Agent Memory
Large-scale evaluation reveals that dedicated memory systems often underperform naive retrieval-augmented generation, introducing a framework of memory benefit and memory tax to assess net value.
Leak Exposes 41,000 Homeland Security Workers–Including Undercover ICE Agents
The leaked records include the locations of ICE and CBP facilities, workers' cellphone numbers, and the members of groups tied to wiretap programs, Palantir software, and xAI’s Grok chatbot.
Jio Platforms, India's largest telecom and digital services company, jumps 21.4% in its trading debut, after raising ~$3.2B in the country's largest-ever IPO
Jio Platforms' shares rose 21.4% on their trading debut Wednesday after a record initial public offering that was lapped up by retail as well as institutional investors.
DLSS-5-Swaper-New/.github
DLSS 5 Swapper is a specialized Windows utility for PC gaming enthusiasts interested in managing compatible graphics components and exploring rendering configurations.
Can Discrete Tokens Serve as Semantically Lossless Representations of Visual Signals?
Factorized Vector Quantization improves discrete visual token quality by using parallel codebooks on complementary feature views, enabling discrete representations to match continuous ones in multimodal understanding.
The Crypto Dream is Structurally Illegal
The initial promise of crypto was digital money without a middleman. Censorship resistance. Financial sovereignty. Anonymity. Satoshi Nakamoto's 2008 whitepaper describes a payment system with no bank, no processor, and no government stand…
Git for the US Code
Lawyers should be able to read a statute as it stood on any date and see which law wrote each phrase. A history of nine titles of the US Code, built by replaying every amendment, as the example.
Artificial Analysis launches the Intelligence Index v5, with new coding and science benchmarks, built on private datasets
Announcing Artificial Analysis Intelligence Index v5, featuring 2 new benchmarks and 3 upgrades. For the first time, the majority of the Index is built on private datasets The Artificial Analysis Intelligence Index is our composite…
HABIT-Bench: Evaluating Habit Induction from Longitudinal Weak Evidence in Agent Memory
HABIT-Bench evaluates whether LLM agent memory systems can infer implicit user habits from sparse dialogue evidence and apply them correctly across long-term interactions.
Ink droplets in motion snag top prize in Nikon Small World contest
Ole Bielfeldt of Germany captured a microscopic view of an inkjet printer depositing ink onto paper.
A Health Update
I’m writing this post from a hospital bed. And I couldn’t be happier. You hear folks say chemotherapy sucks, but until you go through it yourself, you don’t really know how bad it can be. It remains a popular activity only because it’s…
GitHub says Copilot Fusion, which handles model selection and orchestration, is now generally available, releases the GitHub Copilot app, and more
GitHub has debuted GitHub Copilot Fusion, a new system that handles model selection and orchestration for coding tasks so developers don't have to.
Subtly wrong advice misaligns models more than obviously wrong advice - Replication
tl;dr * We fine-tune 12 open-weight models on correct, obviously incorrect, and subtly incorrect advice across 8 narrow domains * We replicate the OpenAI result and find that subtly incorrect data produces more emergent misalignment than o…
CodeCap-Bench: Can Code Serve as a Good Video Caption?
Executable code representations of video content improve reasoning and reconstruction over natural language, yet current multimodal models still struggle to generate faithful code captions for complex temporal dynamics.
Wealth Inequality in America drawn to scale
Picture U.S. wealth as stacks of $100 bills: the median household’s is about 9 inches tall. The average top-1% household’s is 145 feet, and it keeps getting taller and taller…
Research: professional analysts followed AI agents' randomly assigned “leans” on subjective analytical decisions 86% of the time; 77% used the agents' arguments
Amaç Herdağdelen, Hailey Joren, Alex Tamkin — When experts hand their work to an AI agent …
Code-execution flaw in UEFI firmware affects most servers on planet Earth
Critical flaw in EDK2 allows attackers to execute code early in the boot process.
ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning
ThinkWorld improves robotic generalization by representing actions as dynamics chains and using world imagination to learn from unlabeled human videos.
Meta censored the No Kings Instagram account during a nationwide campaign
Meta censored a political organisation during a national voting campaign, blamed its own automation, then claimed its automated systems had detected another automated tool.
Google and Xreal launch the Xreal Aura wired XR glasses, with Android XR and a puck with Snapdragon Reality Elite, for $1,099+ through December, then $1,279+
At the Android Developer Summit on Wednesday, Google's smart glasses partner Xreal launched its new Xreal Aura wired smart glasses …
I spent 12 hours writing an article for LessWrong, calling for a more ‘hippie’ rationalism. Then I didn’t post it. This is why (plus a link to the article, explicitly not located on LessWrong)
TLDR I (and many others in the community) wish to see the universe unfold to its highest potential in the metaphorical palm of an AI as aligned as the most benevolent Buddhist monk – but I (we) feel the community's ontology is too narrow t…
Rollback: Experience Backtracking in World Models for Policy Improvement
Rollback improves robot policies by using world models to rewind to past decisions, explore counterfactual actions, and generate fine-grained supervision, while continuously refining dynamics via real-world feedback.
Dayjob: Tax – evaluating AI agents on 50 professional tax assignments
DAYJOB: Tax evaluates AI agents across 50 realistic tax environments, testing their ability to navigate messy records, exercise professional judgment, and deliver reliable tax work.