Histórias consolidadas

Funcionário da Anthropic confirma que o Claude Code agora tem instruções para não reproduzir propriedade intelectual de jogos — empresa "iterando para garantir [...] o equilíbrio correto"

Anthropic alerta que está muito próximo da capacidade RSI, também conhecida como IA auto-construtora TechCrunch

Anthropic reprime clones de jogos eletrônicos de IA do Claude The Verge

Apresentando o Programa Grid Protectors OpenAI

O principal lobista Chris Lehane alerta os candidatos de que apoiar um moratório de data centers pode ajudar a China a ultrapassar os EUA

O presidente Putin informou aos data centers russos não estatais que eles devem reservar 10 % de sua capacidade total para uso da própria infraestrutura de TI do país.

OpenAI diz que sua IA está ficando melhor em fazer coisas que os humanos fazem, como fazer compras e planejar refeições KRDO

Pesquisadores de cibersegurança descobriram um novo malware infostealer para Windows chamado VOLTA-AUTODOXER (também conhecido como VOLTA) que utiliza um modelo de linguagem grande (LLM) para classificar os dados roubados de uma vítima pelo seu valor de monetização. Como outras famílias típicas de stealer, o VOLTA está equipado para coletar senhas e cookies de navegadores baseados em Chromium; dados de perfis do Outlook e espaços de trabalho do Slack;

Dentro da busca da Anthropic por poder de computação de IA The New York Times

Nvidia is investigating a security incident involving its Brev cloud development platform after detecting unauthorized activity between October 4 and October 20. The incident may have exposed SSH keys used to access customer virtual machines. Nvidia has rotated the affected keys and updated customer instances, but says it is unaware of any improper use

O data center proposto pela Google em Hermantown superou um grande obstáculo de aprovação após o conselho municipal apoiá-lo por unanimidade em 30 de outubro.

A Sony pode estar se preparando para abandonar o mercado de players 4K Blu-ray independentes após varejistas no Japão relatarem que deixaram de esperar novas remessas de seu último modelo restante, o UBP‑X700K. A empresa não confirmou um retiro mundial, e o player continua listado nos EUA. No entanto, a retirada da Sony de outros produtos de mídia física, juntamente com saídas da Oppo, Samsung, LG e Reavon, levanta mais questões sobre o futuro dos players de disco dedicados.

URL do Artigo: https://www.ft.com/content/4d91dbab-6fe4-45ab-beee-d77e091cb37f URL dos Comentários: https://news.ycombinator.com/item?id=50342424 Pontos: 35 # Comentários: 12

Anthropic planeja captar até US$ 3,5 bi no Japão em IPO nos EUA CNN Brasil

Diagnosticar falhas de carga de trabalho de IA distribuída com NVIDIA Attribution | Blog Técnico da NVIDIA NVIDIA Developer

Palestra: A Era da Inteligência Artificial BH Eventos

A startup de um ano agora está avaliada em US$ 300 milhões.

OpenAI faz primeira ação de marketing no Brasil e lança ChatGPT como guia oficial do Lollapalooza Times Brasil | CNBC

Resumo de IA e Tecnologia: Por que a Anthropic não está acalmando as águas turbulentas da IA The Washington Post

URL do Artigo: https://www.regenerator1.com/p/anthropics-growth-just-flattened URL dos Comentários: https://news.ycombinator.com/item?id=50344958 Pontos: 80 # Comentários: 64

URL do Artigo: https://www.techpolicy.press/the-sec-should-block-anthropics-ipo-to-protect-the-public/ URL dos Comentários: https://news.ycombinator.com/item?id=50344919 Pontos: 78 # Comentários: 40

"A inteligência artificial é uma commodity", diz cofundador da Galeria CNN Brasil

OpenAI destaca documentos que apoiam o uso justo para treinamento de IA IPWatchdog

Nvidia exibe sinais de venda com os resultados se aproximando; Nvidia é uma compra agora? Investor's Business Daily

A queda nas ações de semicondutores foi um presente a caminho de 2027 (NASDAQ:SMH) Seeking Alpha

Pensilvânia promulga Lei de Publicidade Enganosa com Inteligência Artificial: Novos requisitos de divulgação para publicidade gerada por IA Consumer Finance Monitor

Anthropic confirmou que adicionou novas restrições ao Claude Code para impedir que usuários recriem jogos eletrônicos existentes. Mas a postagem Developers are downgrading their AI on purpose — and it’s working apareceu primeiro no The New Stack.

OpenAI afirma: IA está ficando cada vez melhor em realizar tarefas humanas CNN Brasil

LifeVal OpenAI

Chatbots de IA estão se tornando conselheiros políticos. Suas respostas podem enganar eleitores NPR

CPI | Turquia abre investigação antitruste sobre as ferramentas de busca com IA do Google PYMNTS.com

Empurrão da fundição da Intel enfrenta teste crítico após queda de 45% nas ações em outubro - CNBC

Anthropic diminui prejuízo trimestral e aumenta receita antes de IPO CNN Brasil

O construtor de sites disse que as demissões fazem parte de uma reestruturação mais ampla da sua organização.

URL do Artigo: https://www.cnbc.com/2026/11/02/oracle-files-to-sell-20-billion-in-new-stock-bonds-to-finance-ai-buildout.html URL dos Comentários: https://news.ycombinator.com/item?id=50342284 Pontos: 47 # Comentários: 17

Governo de Abu Dhabi firma quatro acordos com a OpenAI para estruturar gestão pública totalmente nativa em IA Brasil Inovador

The New York Times está processando a empresa de navegador focada em privacidade Brave por supostamente extrair conteúdo protegido por direitos autorais do veículo e distribuí‑lo ilegalmente a seus próprios clientes, como Anthropic e Perplexity.

Luz e trevas: a batalha decisiva em torno da inteligência artificial Brasil de Fato

Tecnologia de mapeamento de células cancerígenas da Johns Hopkins escolhida para projeto de IA da Anthropic Johns Hopkins University

Ações da Intel despencam após Trump injetar dinheiro dos contribuintes na fabricante de chips - Futurism

General-purpose mobile manipulation requires a combination of capable robot hardware, scalable demonstration collection, expressive whole-body control, and effective policy learning. Yet these components are often developed independently, creating mismatches that limit the capabilities of the complete system. We present Pixel, a co-designed full-stack system for learning general-purpose mobile manipulation. Pixel combines a low-cost wheeled humanoid with human-aligned kinematics and compliant grippers, multiple teleoperation and robot-free human data-collection interfaces, a unified whole-body controller, and a common policy learning pipeline across human and robot demonstrations. A hierarchical learning stack further enables language-conditioned visuomotor control and long-horizon task execution. We evaluate the system across hardware, data collection, and policy learning, and demonstr…

Sensorized human demonstrations offer scalable data for mobile manipulation, but embodiment differences complicate policy learning. We present Pixel-UMI, a pipeline for learning environment-aware whole-body policies from human data. We first introduce Environment-Aware Retargeting (EAR), which combines reconstructed scene geometry with demonstrated head and gripper trajectories to determine robot posture and generate collision-aware joint and base targets. Delay-calibrated action assignment converts these trajectories into actions. However, kinematic differences can leave residual pose errors, creating viewpoint mismatches between human demonstrations and robot observations. We therefore introduce Retargeting-Error-Aware Masking (REAM), which uses these errors during training to reduce reliance on mismatched views at deployment. Finally, because estimated actions cannot fully capture co…

Chain-of-Thought (CoT) generation has substantially improved the performance of Large Language Models (LLMs) on complex reasoning tasks, including code generation, data analysis, and exam-style question answering. Despite these advances, purely neural LLMs continue to struggle with elementary logical reasoning problems and lack the determinism, soundness, and reliability characteristic of symbolic reasoning systems. Conversely, classical symbolic methods such as SAT solving and exact cover guarantee correctness and completeness, but require problems to be expressed in highly specialized formal encodings, limiting their applicability to natural language inputs. In this work, we present a synergistic neuro-symbolic framework that bridges this gap by combining neural semantic parsing with deterministic exact cover reasoning. Given a Zebra-style logic puzzle in natural language, an LLM tran…

GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-ste…

Muon has recently emerged as a promising optimizer for pretraining large language models, offering a speedup over AdamW in minimizing pretraining loss. However, it remains unclear whether Muon's advantage persists after fine-tuning. In this paper, we study the downstream behavior of Muon and AdamW and uncover a surprising trend: beyond a sufficiently large pretraining data budget, AdamW outperforms Muon after fine-tuning. Moreover, as model size increases, AdamW overtakes Muon at progressively smaller pretraining token-parameter ratios. To explain this reversal, we view fine-tuning as a perturbation of the outputs and introduce the margin-to-spread ratio, which measures how far the correct logit lies above the wrong ones relative to how widely the wrong logits are spread. In language models, Muon has the larger margin-to-spread ratio at small data budgets and AdamW at large ones, mirror…

Reinforcement learning with verifiable rewards (RLVR) can enable generalizable improvement for large language model (LLM) reasoning. This paper studies how to incentivize such generalization via reasoning strategies. We study this question with a controllable equation-solving framework: finding multiple inputs x for equations f(x)=y, where f is a non-bijective string function. This setting makes reasoning strategies observable while supporting a clear definition of easy-to-hard out-of-distribution (OOD) shifts, e.g., more solutions, longer y, etc. We uncover a subtle RLVR failure: it teaches the model to solve in-distribution tasks with high accuracy and structured step-by-step derivation, yet under OOD shifts, the same model abandons derivation and collapses to ineffective brute-force enumeration, causing sharp accuracy drops. This observation reveals the deeper challenges of RLVR gene…

Group-based reinforcement learning with verifiable rewards (RLVR) learns only from prompts whose sampled completions disagree: all-pass and all-fail groups produce zero relative advantage and burn generation budget without moving the policy. Adaptive schedulers try to steer rollouts toward informative prompts, yet they rely on per-prompt pass-rate trackers that entangle two quantities evolving at very different speeds---the policy's shifting competence across reasoning skills and each prompt's intrinsic difficulty. When a single exponential average or Beta counter tracks both at once, short memory scrambles the relative ordering of prompts whereas long memory reports stale pass rates. We introduce FADE (Factored Ability--Difficulty Estimation), a bilinear state-space filter that separates the two timescales. FADE represents each prompt's success logit as the inner product of a whitened …

Communication is important for our well-being, but it can be impaired by disabilities, such as locked-in syndrome (LIS), cerebral palsy (CP), and dysarthria. Non-invasive brain-to-text brain-computer interfaces (BCIs) offer a promising solution by providing an alternative communication pathway for people with communication disabilities. However, most recent breakthroughs have relied on large, cumbersome, and costly neuroimaging modalities, such as functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG). Numerous studies have attempted decoding using the lighter and cheaper electroencephalography (EEG), but the results remain inferior compared to the other modalities. More importantly, we do not understand how these modalities are related to text at the letter, word, sentence, and higher semantic levels. Many prior studies have focused on decoding below the sentence…

Vision-Language Models (VLMs) hold promise for zero-shot video task-progress estimation in robot learning, yet existing frameworks frequently suffer from confirmation bias, hallucinating progress in failed executions due to implicit success assumptions or relative temporal ranking designs. Naive prompting of frontier models collapses into a binary success detector that fails to capture nuanced, non-linear progress and failures. While post-training approaches can mitigate this through explicit training, they demand expensive real-world robotics data and require computationally intensive post-training for every individual base model. To resolve these limitations, we introduce D&A (Describe & Assess): our approach decouples progress estimation into an objective, task-grounded textual description of scene dynamics followed by an assessment phase using explicit rubrics. By compelling the VLM…

A model is only half of an AI agent system. The other half is the inference strategy: the reasoning procedure that turns model calls into a response. Today that second half is written by hand, and the experience of building one strategy never carries to the next. We introduce Strategist, a meta-agent that evolves inference strategies across tasks. Each strategy is written as a blueprint,executable code that other strategies can reuse, and together the blueprints form a library. Given a task and a small development set, Strategist retrieves related blueprints, writes candidate strategies, tests them on real samples, and revises them from the results. Every candidate then enters the library, so later tasks build on earlier discoveries. On 30 benchmarks, Strategist beats the strongest baseline on 24 and matches its average accuracy (74.7% vs.74.0%) at 39% lower inference cost. Its strategi…

Large language models (LLMs) show growing promise for scientific discovery, yet basic questions remain about what enables reliable discovery. How can we trace the scientific discovery capacity of LLMs to their training data? We study this question as one of missing probability mass and develop a theoretical account in which monofacts-the facts observed once during training-set the discovery capacity. In materials science and regulatory genomics, we directly manipulate monofact rates in training sets and find that more monofacts consistently yield more discoveries across two open-weight model families, sampling strategies, and reasoning budgets. Further experiments show the relationship cannot be explained by training-set diversity alone and can transfer across scientific domains when paired with in-domain data. Together, our results identify monofact rate in training data as a key facto…

Evaluating a generative model on all queries from all available benchmark tasks is prohibitively expensive. As such, models are often evaluated on different subsets of tasks and/or different subsets of queries from a given task. The resulting collection of responses and the corresponding (model, task) score matrix are sparse and unpaired. In this paper, we introduce the Product Kernel Perspective Space (PKPS), a collection of low-dimensional representations of models based on their embedded responses. PKPS generalizes existing methods for representing black-box generative models by comparing responses to similar -- rather than just identical -- queries and is therefore able to utilize sparse and unpaired information. Theoretically, we show that benchmark prediction using PKPS is query-efficient relative to methods that only use information from dense and paired portions of the evaluatio…

Os dias de apontar um agente de IA para um backlog e deixá-lo consumir recursos sem muita supervisão podem O post Seu agente de codificação de IA agora funciona com orçamento — e pode interromper você no meio da tarefa apareceu primeiro no The New Stack.

URL do Artigo: https://terrytao.wordpress.com/2026/11/02/egoless-mathematics-in-the-ai-era/ URL dos Comentários: https://news.ycombinator.com/item?id=50341736 Pontos: 44 # Comentários: 37

O que acontece quando eleitores recorrem a chatbots de IA para obter informações eleitorais? MPR News

Ações da Nvidia parecem surpreendentemente baratas. Por que NVDA pode subir 61%. Barchart.com

Increased computing power has transformed computational science and improved predictive capabilities for complex systems, but many phenomena still exceed current simulation capabilities or require rapid evaluations for fast design iterations. Closure models address this challenge by approximating unresolved effects in reduced or incomplete dynamical systems. Data-driven closures offer a flexible mechanism for improving model accuracy and applicability, but approaches that embed the closure within a computational model for construction and training phases can be computationally expensive and cause numerical instabilities in the computational model during training. To avoid the challenges posed by embedded approaches, staged approaches to closure-model learning have been developed. The first stage uses data assimilation to infer state and closure trajectories from observations. The second…

ufakzeka-kulak is an open Turkish speech-to-text model of 31.6M parameters, a FastConformer encoder with a token-and-duration transducer (TDT) and a CTC head, trained from scratch and released under Apache-2.0. It writes cased, punctuated Turkish with numbers as words. Its 4,977 training hours come from sources whose licences allow commercial use; 281 hours are human-transcribed, and for the other 4,696 the training text is an uploader subtitle or a machine transcript, kept only where a second, independent reading of the audio agrees with it. The recipe, checkpoint, labelling and export choices were made on development data, and the test sets were scored once, in a release gate fixed before training. With greedy TDT decoding the normalised word error rate (WER) is 0.0921 (95 percent interval 0.0792 to 0.1055) on FLEURS Turkish test and 0.0916 (0.0873 to 0.0958) on Common Voice 27.0 Turk…

As pretraining scales, medical language models have advanced rapidly in clinical question answering. Their ability to achieve effective physiological control through sustained intervention still lacks systematic quantitative evaluation. Type 1 diabetes provides a concrete setting: impaired insulin secretion creates dependence on exogenous insulin, with dosing continually adjusted using information about meals, exercise, illness, and glucose measurements. We introduce DiabAgent, a research suite for agentic physiological control comprising the Agentic Physiological Control (APC) framework, the DiabetSim simulator, and the SweetBench benchmark. APC controls glucose through medical-device tools, using ongoing observations to infer individual states, accumulate personal experience, and choose treatment actions. We build DiabetSim to simulate different patients, events, and device conditions…

AI agents are increasingly deployed in settings where they interact and coordinate with other agents, yet honesty is mostly measured in single-turn, single-agent benchmarks that pressure the model to deceive. We study deception that emerges in multi-agent interaction without any instruction to deceive. We introduce Kradle, a platform for evaluating AI models, and Four Bridges, an information asymmetry game in which four agents distribute across four rooms: one is the death room, the others hold food worth more if not shared. One agent, the informed agent, knows which room is the death room and can withhold or disclose that information. We evaluate eight frontier models as the informed agent. Deceptive behavior appears in every model and in 73% of all runs (6% to 100% per model), and does not decrease with capability tier within a model family. When the informed agent does disclose, the …