News from October 27, 2026
In my view, the accusation against the Trump administration and the Anthropic ruling show how technology carries choices about transparency and using others’ work.
The day in three lines
- Two senators accuse the Trump administration of breaking the law by withholding a court ruling on surveillance.
- A federal judge ruled that Anthropic pirated Stanford professors’ books to train its systems.
- Google announced Genie 4, which creates interactive simulations with visuals, sounds, and dialogue and lets users change scenes.
-
Ghostty 1.4
Ghostty 1.4 improves speed and memory use, displays animated images, and adds compatibility updates and improvements for remote access.
Quote
Release notes for Ghostty 1.4.0, released on October 27, 2026.
-
How to change your perception of time (and make life seem longer)
BBC Science Focus asks whether changing how time feels can make life seem longer, but the text provides no explanation.
Quote
Can we feel like we have more time by bending our perception of it?
-
Your Robot’s Safety Functions Already Work. What If the Input Lies?
Tests show that manipulated text, images, sounds, or messages can change robots’ behavior, and researchers urge teams to assess the resulting safety risks.
Quote
This article is brought to you by VicOne. A mobile robot slows for someone in a hallway. A collaborative arm eases its movement as a worker approaches. A humanoid pauses to let a person pass.
-
Retell uses OpenAI models to help power 100M monthly calls
Retell says an OpenAI model led its voice test, while a coding tool helped increase revenue per employee.
Quote
GPT-6 Astra tops Retell’s voice agent benchmark, while Codex helps it reach a $1.6M annual revenue run rate per employee.
-
O poder da matemática a serviço da indústria
CeMEAI brought universities and companies together in Brazil, applying mathematics to industrial problems and supporting research on proteins.
Quote
Centro de referência em matemática aplicada ajudou a aproximar academia e indústria no Brasil
-
Work Agents: Systems, Capabilities, and Optimization
A paper proposes a framework for understanding and evaluating digital agents that perform work tasks, and identifies capabilities needed for reliable results.
Quote
As foundation models evolve from information generation to digital work execution, agents are increasingly capable of completing complex tasks across tools, applications, and workflows.
-
Unknown bald eagle spotted on Shadow’s nest
Shadow has been spending less time at the nest he once shared with Jackie.
Quote
Shadow has been spending less time on the nest he once shared with Jackie.
-
I Used AI to Fake a Scientific Paper. Here’s What Happened
The author used free artificial intelligence tools to create a fake scientific paper about light pulses that looked convincing.
Quote
Science has a spam problem. And the spam filter is getting clogged up. According to all reports, academic publishing is drowning in what has come to be called AI slop: low-quality, mass-produced content generated with AI.
-
Build specialized multimodal models with LFM3
Liquid AI introduced models for text, images, and audio that users can tailor with their own data and run on local devices.
Quote
A Blog post by Liquid AI on Hugging Face
-
O Impa na China
The author arrived in Beijing for a work visit and reflected on how China had changed over three decades.
Quote
Após um voo de mais de 24 h, pousei no aeroporto internacional de Pequim às 6h da manhã do último dia 19, uma segunda-feira, para mais uma visita de trabalho à China.
-
US Government is hiding a secret court ruling about how it spies on Americans
Two senators say the Trump administration is breaking the law by withholding a court ruling on surveillance while asking Congress to renew the program.
Quote
Two US senators say the Trump administration is breaking the law by withholding the FISA court decision while it presses Congress to renew America’s most sweeping surveillance program.
-
How brain implants turn thoughts into speech
A brain implant captures Casey Harrell’s signals for intended speech and turns them into words a computer can speak in a voice resembling his own.
Quote
New brain-computer interfaces turn attempted speech into words and can speak them in a voice resembling the user’s own.
-
OpenAI expands partnership with Oracle to bring its apps to ChatGPT
OpenAI and Oracle expanded their partnership to bring business apps and AI assistants to ChatGPT and develop new Oracle offerings.
Quote
OpenAI and Oracle are expanding their partnership to bring business applications and AI agents into ChatGPT and power new Oracle offerings.
-
Autenticação digital sem fricção: como validar identidades sem complicar a jornada
Companies can combine identity, device data, and behavior to request extra checks when an operation seems riskier.
Quote
Entenda como identidade, contexto e risco ajudam a definir quando uma verificação adicional é necessária
-
StreamVibe: Proactive Video Agents for Evolving User Intent
A study presents a system that follows videos and tries to infer what a person needs, even without explicit requests, adjusting responses as situations change.
Quote
Vision-language models are shifting from offline passive understanding toward online proactive interaction.
-
The Vera C. Rubin Observatory may have just discovered our solar system’s sixth dwarf planet
Astronomers found a distant object beyond Pluto, but further observations must confirm whether it is round enough to qualify as a dwarf planet.
Quote
Located farther away from Earth than Pluto, the object was hidden in the light of the Milky Way’s stars
-
Matemático Terence Tao diz estar em estado de guerra após anúncios da OpenAI
Terence Tao said he was in a state of war at a gathering where mathematicians discussed how artificial intelligence is affecting their field.
Quote
"Estou em estado de guerra", disse Terence Tao, 51, que costuma ser descrito como um dos maiores matemáticos do mundo.
-
Genie 4
Google announced a system that creates interactive simulations with visuals, sounds, and dialogue, letting users change scenes in real time.
-
Origami-based construction method yields lightweight, load-bearing structures on architectural scales
MIT researchers built lightweight, strong structures from folded pieces that use less material and can be assembled without interior columns.
Quote
The modular approach could be used to erect emergency shelters or fabricate aerospace components more quickly and with less material than some techniques.
-
Trump diz que vai proibir o termo 'inteligência artificial'
Trump said he plans to ban “artificial intelligence” in official documents and use “superintelligence” instead, but gave no details on how or when.
Quote
Presidente dos EUA considera o termo impreciso e defende que a tecnologia passe a ser chamada de ‘superinteligência’; declaração foi publicada na rede Truth Social.
-
InteractiveWorld: A Verifiable Human-Centric Environment for Building Better Interactive MLLM Assistants
A study created a virtual environment to test AI assistants guiding people through physical tasks and found they still often fail.
Quote
People have long dreamed of an assistant that is always there to guide them and remind what they forget, from new dishes to lost keys.
-
8 simple, surprising ways to improve your sleep, according to world-leading experts
The text says better sleep depends on falling asleep faster and promises eight research-backed tips, but does not provide them.
Quote
The secret to sleeping longer isn’t spending more time in bed, it’s falling asleep faster
-
Chimpanzés são menos propensos a roubar comida de amigos
Pasa and Yoyo, two chimpanzee friends at a sanctuary in Uganda, rest together, care for each other, and share food.
Quote
Pasa e Yoyo são melhores amigas. Elas passam as tardes descansando juntas na grama do Santuário de Chimpanzés da Ilha de Ngamba, em Uganda, cuidando uma da outra e compartilhando comida. Às vezes, até tentam fugir juntas do local.
-
AI promised less work. I've never worked more
The author says AI agents made coding cheaper, leading him to take on more tasks at once and feel more mentally drained.
Quote
Agents made implementation cheap, so I started doing far more of it in parallel.
-
Why are bogs so good at preserving dead bodies?
No summary: the source’s text could not be read.
-
Empresas recorrem à religião para ensinar ética à IA
Companies consult religious leaders and philosophers to guide artificial intelligence ethics, but experts warn that religions cannot be reduced to rules.
Quote
Gigantes da tecnologia consultam líderes religiosos e filósofos em busca de princípios para orientar seus modelos, mas especialistas alertam para os limites dessa abordagem.
-
Spatial Audio Generation and Editing with Executable Plans and Self-Distilling Transfusion
A study introduces a system that creates and edits spatial audio using plans describing sounds, movement, and elements to preserve.
Quote
Spatial audio authoring coordinates source decisions with soundfield synthesis. For editing, plans specify what to change and retain, while reference audio supplies further acoustic detail.
-
OpenAI just scooped human researchers on a million-dollar quantum math problem
OpenAI used artificial intelligence to present a mathematical proof about the theory describing particles, preceding the work of two mathematicians.
Quote
Another bombshell result shows how artificial intelligence is fueling fierce competition at the frontiers of math and physics
-
Celular com Android: saiba configurações para comprar um aparelho que dura mais e não trava
A guide recommends Android phones with ample memory, storage and battery, a smooth screen, and several years of updates.
Quote
Lista reúne os principais pontos para quem busca bom desempenho, mais espaço e maior vida útil no dia a dia.
-
Anthropic trained Claude on illegally obtained Stanford professors' books
A federal judge ruled that Anthropic pirated Stanford professors’ books to train its systems; some authors filed separate lawsuits after a $1.5 billion settlement.
Quote
The AI company pirated at least 10 Stanford academics’ work in a $1.5 billion copyright settlement.
14 more
-
The immune system reboot: How to prime your body’s defences to slow ageing
Ignore pseudo-science supplements promising to solve inflammation & prevent illness, and learn the truth about recalibrating your body’s defences for long-term health
-
WorldCast: World-Consistent Character Video Generation with Large-Angle Camera Control
A study presents a method for generating character videos that follow wide camera movements while preserving the character’s appearance and identity.
Quote
Precise camera control is essential for cinematic character video generation, yet existing methods struggle to follow large-angle trajectories with combined rotation, translation, and radius variation from arbitrary initial azimuths.
-
China’s super-secret space plane visible in clearest photo yet
The U.S. quietly included a picture in a recent unclassified report.
Quote
The U.S. quietly dropped the picture inside a recent unclassified report.
-
Python's experimental JIT to be removed
The maintainers withdrew their proposal to keep Python’s experimental compiler, and development of the project appears to have ended.
Quote
Python's experimental just-in-time (JIT) compiler, which was added to CPython in 2024 for the 3
-
How a 16th century pirate accidentally spawned a centuries-old math problem
A question from Walter Raleigh about cannonballs led mathematicians to investigate for centuries how to stack spheres with the least empty space.
Quote
A question posed by Sir Walter Raleigh about cannonballs inadvertently led to a puzzle that mathematicians have been trying to solve for centuries
-
Trajectory2Gym: Decision-Grounded Compositional Interaction Synthesis for Tool-Use Agents
A study introduces a way to create training interactions for tool-using agents, grounding decisions in visible history, and reports better test performance.
Quote
Successful tool-use trajectories are not necessarily useful supervision.
-
Survival LLM: AI assistant and emergency library that runs on your own computer
A free emergency assistant works offline on your computer and shows the sources it uses to answer questions.
Quote
A free AI assistant and survival library (1.5 million offline articles) that runs on your own computer, with no internet. Shows the source of every answer. Mac and Windows.
-
DynaFlow: Learning Free-Form Multi-Agent Orchestration
A study presents an agent that creates and adjusts collaboration steps among other agents, outperforming a comparable system across six tests.
Quote
Multi-agent systems often rely on predefined collaboration structures, limiting their ability to adapt to diverse task requirements.
-
Wikipedia's Cleanup of AI-Generated Content
AI-generated text is harder to spot, but that hasn’t deterred them
-
Eat Your Precision: Precision Objective Mitigates Reward Hacking in Rubric-based RL
A study presents a training method that makes AI models’ answers more concise and accurate while preserving coverage of the criteria.
Quote
Rubric-based rewards enable effective reinforcement learning on non-verifiable tasks.
-
The Mythical Man-Month in the Age of AI
The author argues that AI agents do not eliminate software project problems: without clear instructions, they can produce wrong results faster.
Quote
An update to Brooks's Law in the age of AI.
-
Continual Learning Exam: Benchmarking Long-Horizon RSI and Experiential Learning in LLM Agents
A benchmark of professional tasks found that feedback improves AI agents’ performance, but they still struggle to learn from experience over time.
Quote
Continual learning is the ability of LLM agents to improve over time by acquiring, refining, and applying knowledge from experience while retaining useful prior knowledge.
-
Roblox downloaded 145GB models from us (including adult content) without asking
Open3DLab says a crawler identifying itself as Roblox downloaded models, including adult content, without permission, possibly to train artificial intelligence.
Quote
On 15 and 16 October 2026, a crawler that said it belonged to Roblox walked through almost every project on Open3DLab, SFMLab and SmutBase, and proceeded to download complete copies of 1,027 files from 712 projects. 145.5 GB of our…
-
Trial by Trace: Benchmarking AI Agent Harnesses for Reliability and Failure Attribution
A study presents a way to assess the reliability of systems supporting artificial intelligence agents and identify failures, including in successfully completed tasks.
Quote
LLM agents are increasingly used to solve real-world problems, and their effectiveness depends on both the underlying models and the harnesses that support their operation.
Saved
Kept only in this browser.
Nothing saved yet. Tap Save on any link to read it later.
How this is made
A script picks the links by votes or by date. An AI summarizes each one using only the source’s text, and writes the day’s three lines from those summaries. Nobody reviews them before they go up, so check the original before citing.
