https://aiexpert.news/en/article/openai-publishes-our-principles-what-enterprise-compliance-teams-should-parse ai|expert en 2026-04-27T00:44:15.509Z OpenAI Publishes "Our Principles" — What Enterprise Compliance Teams Should Parse https://aiexpert.news/pt/article/openai-publishes-our-principles-what-enterprise-compliance-teams-should-parse ai|expert pt-BR 2026-04-27T00:44:15.509Z OpenAI publica "Nossos Princípios" — O que as equipes de conformidade corporativa devem analisar https://aiexpert.news/es/article/openai-publishes-our-principles-what-enterprise-compliance-teams-should-parse ai|expert es 2026-04-27T00:44:15.509Z OpenAI publica "Nuestros Principios" — Lo que los equipos de cumplimiento empresarial deben analizar https://aiexpert.news/en/article/federal-register-rfc-puts-ai-assisted-adjudication-of-government-applications-on ai|expert en 2026-04-27T00:20:15.372Z ACUS RFC on Frontline AI Adjudication Names SSA, USPTO, VA, SBA, and DOL https://aiexpert.news/pt/article/federal-register-rfc-puts-ai-assisted-adjudication-of-government-applications-on ai|expert pt-BR 2026-04-27T00:20:15.372Z RFC da ACUS sobre Adjudicação por IA na Linha de Frente Cita SSA, USPTO, VA, SBA e DOL https://aiexpert.news/es/article/federal-register-rfc-puts-ai-assisted-adjudication-of-government-applications-on ai|expert es 2026-04-27T00:20:15.372Z RFC de ACUS sobre Adjudicación por IA en Primera Línea Nombra SSA, USPTO, VA, SBA y DOL https://aiexpert.news/en/article/anthropic-equity-is-now-a-bay-area-real-estate-currency-what-that-signals ai|expert en 2026-04-27T00:00:15.344Z Mill Valley Seller Lists 13-Acre Estate for Anthropic Equity, Not Cash https://aiexpert.news/pt/article/anthropic-equity-is-now-a-bay-area-real-estate-currency-what-that-signals ai|expert pt-BR 2026-04-27T00:00:15.344Z Vendedor em Mill Valley Aceita Equity da Anthropic — Não Dinheiro — por Propriedade de 13 Acres https://aiexpert.news/es/article/anthropic-equity-is-now-a-bay-area-real-estate-currency-what-that-signals ai|expert es 2026-04-27T00:00:15.344Z Vendedor en Mill Valley Exige Equity de Anthropic — No Efectivo — por Propiedad de 13 Acres https://aiexpert.news/en/article/new-research-pins-minimax-sample-complexity-of-multicalibration-a-key-fairness-g ai|expert en 2026-04-26T22:30:15.332Z Multicalibration at 1% Error Demands One Million Training Samples, Researchers Prove https://aiexpert.news/pt/article/new-research-pins-minimax-sample-complexity-of-multicalibration-a-key-fairness-g ai|expert pt-BR 2026-04-26T22:30:15.332Z Multicalibração com 1% de Erro Exige Um Milhão de Amostras de Treinamento, Provam Pesquisadores https://aiexpert.news/es/article/new-research-pins-minimax-sample-complexity-of-multicalibration-a-key-fairness-g ai|expert es 2026-04-26T22:30:15.332Z La Multicalibración con 1% de Error Exige Un Millón de Muestras de Entrenamiento, Prueban Investigadores https://aiexpert.news/en/article/openais-chatgpt-ad-targeting-goes-deeper-than-users-realize-conversations-are-th ai|expert en 2026-04-26T22:14:15.295Z OpenAI's ChatGPT Ad Targeting Goes Deeper Than Users Realize — Conversations Are the Data https://aiexpert.news/pt/article/openais-chatgpt-ad-targeting-goes-deeper-than-users-realize-conversations-are-th ai|expert pt-BR 2026-04-26T22:14:15.295Z O Targeting de Anúncios do ChatGPT da OpenAI Vai Mais Fundo do que os Usuários Percebem — as Conversas São os Dados https://aiexpert.news/es/article/openais-chatgpt-ad-targeting-goes-deeper-than-users-realize-conversations-are-th ai|expert es 2026-04-26T22:14:15.295Z El Targeting de Anuncios de ChatGPT de OpenAI Va Más Profundo de lo que los Usuarios Creen — las Conversaciones Son los Datos https://aiexpert.news/en/article/agentic-ai-framework-auto-converts-research-questions-into-executable-scientific ai|expert en 2026-04-26T21:56:15.270Z Agentic Framework Hits 83% Intent Accuracy by Confining LLM to Query Parsing https://aiexpert.news/pt/article/agentic-ai-framework-auto-converts-research-questions-into-executable-scientific ai|expert pt-BR 2026-04-26T21:56:15.270Z Framework Agêntico Atinge 83% de Acurácia de Intenção ao Confinar o LLM ao Parsing de Consultas https://aiexpert.news/es/article/agentic-ai-framework-auto-converts-research-questions-into-executable-scientific ai|expert es 2026-04-26T21:56:15.270Z Framework Agéntico Alcanza 83% de Precisión de Intención al Confinar el LLM al Parsing de Consultas https://aiexpert.news/en/article/deep-learning-cuts-power-grid-milp-solve-time-with-warm-start-framework ai|expert en 2026-04-26T21:20:15.320Z Transformer Warm-Start Hits 100% Feasibility and Cuts Cost in 20% of Grid Runs https://aiexpert.news/pt/article/deep-learning-cuts-power-grid-milp-solve-time-with-warm-start-framework ai|expert pt-BR 2026-04-26T21:20:15.320Z Warm-Start com Transformer Atinge 100% de Viabilidade e Reduz Custo em 20% das Execuções na Rede Elétrica https://aiexpert.news/es/article/deep-learning-cuts-power-grid-milp-solve-time-with-warm-start-framework ai|expert es 2026-04-26T21:20:15.320Z El Warm-Start con Transformer Alcanza 100% de Viabilidad y Reduce el Costo en el 20% de las Ejecuciones en la Red Eléctrica https://aiexpert.news/en/article/giva-challenges-loras-dominance-with-gradient-seeded-vector-fine-tuning ai|expert en 2026-04-26T21:08:15.250Z GiVA cuts vector fine-tuning rank 8-fold to match LoRA training speed https://aiexpert.news/pt/article/giva-challenges-loras-dominance-with-gradient-seeded-vector-fine-tuning ai|expert pt-BR 2026-04-26T21:08:15.250Z GiVA reduz o rank do fine-tuning vetorial em 8× para igualar a velocidade de treino do LoRA https://aiexpert.news/es/article/giva-challenges-loras-dominance-with-gradient-seeded-vector-fine-tuning ai|expert es 2026-04-26T21:08:15.250Z GiVA reduce el rango del fine-tuning vectorial en 8× para igualar la velocidad de entrenamiento de LoRA https://aiexpert.news/en/article/lora-redux-new-survey-settles-which-fine-tuning-variant-to-use-and-when ai|expert en 2026-04-26T20:54:15.322Z New LoRA Survey Replaces Fine-Tuning Folklore With Signal-Processing Criteria https://aiexpert.news/pt/article/lora-redux-new-survey-settles-which-fine-tuning-variant-to-use-and-when ai|expert pt-BR 2026-04-26T20:54:15.322Z Nova Pesquisa sobre LoRA Substitui o Folclore do Fine-Tuning por Critérios de Processamento de Sinais https://aiexpert.news/es/article/lora-redux-new-survey-settles-which-fine-tuning-variant-to-use-and-when ai|expert es 2026-04-26T20:54:15.322Z Nueva Investigación sobre LoRA Reemplaza el Folclore del Fine-Tuning con Criterios de Procesamiento de Señales https://aiexpert.news/en/article/halluscope-reveals-how-crafted-prompts-override-vision-in-multimodal-llms ai|expert en 2026-04-26T20:40:15.238Z False Prompt Assumptions Outrank Vision Failures in New LVLM Hallucination Study https://aiexpert.news/pt/article/halluscope-reveals-how-crafted-prompts-override-vision-in-multimodal-llms ai|expert pt-BR 2026-04-26T20:40:15.238Z Falsas Suposições no Prompt Superam Falhas de Visão em Novo Estudo de Alucinação em LVLMs https://aiexpert.news/es/article/halluscope-reveals-how-crafted-prompts-override-vision-in-multimodal-llms ai|expert es 2026-04-26T20:40:15.238Z Las Falsas Suposiciones en el Prompt Superan los Fallos de Visión en Nuevo Estudio de Alucinación en LVLMs https://aiexpert.news/en/article/openai-warns-developers-old-prompts-actively-hurt-gpt-55-performance ai|expert en 2026-04-26T20:28:15.320Z OpenAI Says Legacy Prompts Degrade GPT-5.5 Output, Urges Full Rebuild https://aiexpert.news/pt/article/openai-warns-developers-old-prompts-actively-hurt-gpt-55-performance ai|expert pt-BR 2026-04-26T20:28:15.320Z OpenAI Diz que Prompts Legados Degradam o GPT-5.5 e Exige Reconstrução Total https://aiexpert.news/es/article/openai-warns-developers-old-prompts-actively-hurt-gpt-55-performance ai|expert es 2026-04-26T20:28:15.320Z OpenAI Advierte que los Prompts Heredados Degradan el GPT-5.5 y Exige una Reconstrucción Total https://aiexpert.news/en/article/mathduels-self-play-benchmark-cracks-the-frontier-model-ceiling-problem ai|expert en 2026-04-26T20:15:46.460Z Tested on 19 Frontier Models, MathDuels Decouples Authoring From Solving Skill https://aiexpert.news/pt/article/mathduels-self-play-benchmark-cracks-the-frontier-model-ceiling-problem ai|expert pt-BR 2026-04-26T20:15:46.460Z Testado em 19 Frontier Models, MathDuels Separa Habilidade de Autoria da de Resolução https://aiexpert.news/es/article/mathduels-self-play-benchmark-cracks-the-frontier-model-ceiling-problem ai|expert es 2026-04-26T20:15:46.460Z Probado en 19 Frontier Models, MathDuels Desacopla la Habilidad de Autoría de la de Resolución https://aiexpert.news/en/article/openai-retires-standalone-codex-absorbs-coding-capability-into-gpt-55 ai|expert en 2026-04-26T20:02:33.233Z OpenAI Folds Codex Into GPT-5.5, Forcing Enterprise Migration at 20% Price Hike https://aiexpert.news/pt/article/openai-retires-standalone-codex-absorbs-coding-capability-into-gpt-55 ai|expert pt-BR 2026-04-26T20:02:33.233Z OpenAI Incorpora o Codex ao GPT-5.5, Forçando Migração Empresarial com Alta de 20% no Preço https://aiexpert.news/es/article/openai-retires-standalone-codex-absorbs-coding-capability-into-gpt-55 ai|expert es 2026-04-26T20:02:33.233Z OpenAI Fusiona Codex en GPT-5.5, Forzando Migración Empresarial con un Alza de Precio del 20% https://aiexpert.news/en/article/anthropic-builds-test-marketplace-for-agent-on-agent-commerce ai|expert en 2026-04-26T19:48:33.192Z Anthropic Finds Agent Quality Gaps Skew Outcomes in Peer-to-Peer Markets https://aiexpert.news/pt/article/anthropic-builds-test-marketplace-for-agent-on-agent-commerce ai|expert pt-BR 2026-04-26T19:48:33.192Z Anthropic Identifica Lacunas de Qualidade em Agentes que Distorcem Resultados em Mercados Peer-to-Peer https://aiexpert.news/es/article/anthropic-builds-test-marketplace-for-agent-on-agent-commerce ai|expert es 2026-04-26T19:48:33.192Z Anthropic Detecta Brechas de Calidad en Agentes que Distorsionan Resultados en Mercados Peer-to-Peer https://aiexpert.news/en/article/cambridges-new-brain-inspired-memristor-claims-70-cut-in-ai-energy-draw-and-its- ai|expert en 2026-04-26T04:48:17.348Z Cambridge Hafnium-Oxide Memristor Targets 70% Cut in AI System Energy https://aiexpert.news/pt/article/cambridges-new-brain-inspired-memristor-claims-70-cut-in-ai-energy-draw-and-its- ai|expert pt-BR 2026-04-26T04:48:17.348Z Memristor de Óxido de Háfnio de Cambridge Mira Redução de 70% no Consumo de Energia de Sistemas de IA https://aiexpert.news/es/article/cambridges-new-brain-inspired-memristor-claims-70-cut-in-ai-energy-draw-and-its- ai|expert es 2026-04-26T04:48:17.348Z Memristor de Óxido de Hafnio de Cambridge Apunta a un Recorte del 70% en la Energía de Sistemas de IA https://aiexpert.news/en/article/hesses-glass-bead-game-was-a-warning-llms-built-it-anyway ai|expert en 2026-04-25T23:50:21.722Z LLMs Have Instantiated Hesse's Glass Bead Game, Boymal Argues https://aiexpert.news/pt/article/hesses-glass-bead-game-was-a-warning-llms-built-it-anyway ai|expert pt-BR 2026-04-25T23:50:21.722Z LLMs Instanciaram o Jogo de Contas de Vidro de Hesse, Argumenta Boymal https://aiexpert.news/es/article/hesses-glass-bead-game-was-a-warning-llms-built-it-anyway ai|expert es 2026-04-25T23:50:21.722Z Los LLMs Han Instanciado el Juego de los Abalorios de Hesse, Argumenta Boymal https://aiexpert.news/en/article/africas-data-labelers-are-organizing-against-the-ai-giants-profiting-from-their- ai|expert en 2026-04-25T23:36:01.136Z Kenya's Data Labelers Association Targets Apple, Meta and Google Over Wages and NDAs https://aiexpert.news/pt/article/africas-data-labelers-are-organizing-against-the-ai-giants-profiting-from-their- ai|expert pt-BR 2026-04-25T23:36:01.136Z Associação de Rotuladores de Dados do Quênia Mira Apple, Meta e Google por Salários e NDAs https://aiexpert.news/es/article/africas-data-labelers-are-organizing-against-the-ai-giants-profiting-from-their- ai|expert es 2026-04-25T23:36:01.136Z La Asociación de Etiquetadores de Datos de Kenia Apunta a Apple, Meta y Google por Salarios y NDAs https://aiexpert.news/en/article/yann-lecuns-ami-labs-raises-1b-to-build-modular-ai-that-rejects-the-llm-playbook ai|expert en 2026-04-25T17:18:39.807Z Yann LeCun's AMI Labs raises $1B on a thesis that modular AI beats LLMs https://aiexpert.news/pt/article/yann-lecuns-ami-labs-raises-1b-to-build-modular-ai-that-rejects-the-llm-playbook ai|expert pt-BR 2026-04-25T17:18:39.807Z AMI Labs, de Yann LeCun, capta $1B com a tese de que IA modular supera os LLMs https://aiexpert.news/es/article/yann-lecuns-ami-labs-raises-1b-to-build-modular-ai-that-rejects-the-llm-playbook ai|expert es 2026-04-25T17:18:39.807Z AMI Labs, de Yann LeCun, recauda $1B con la tesis de que la IA modular supera a los LLMs https://aiexpert.news/en/article/deepminds-aletheia-solves-6-of-10-unpublished-math-problems-autonomously-and-ref ai|expert en 2026-04-25T16:44:28.763Z DeepMind Aletheia Solves 6 of 10 Research Math Problems, Refuses to Fake the Others https://aiexpert.news/pt/article/deepminds-aletheia-solves-6-of-10-unpublished-math-problems-autonomously-and-ref ai|expert pt-BR 2026-04-25T16:44:28.763Z Aletheia da DeepMind Resolve 6 de 10 Problemas Matemáticos de Pesquisa e Recusa Falsificar os Outros https://aiexpert.news/es/article/deepminds-aletheia-solves-6-of-10-unpublished-math-problems-autonomously-and-ref ai|expert es 2026-04-25T16:44:28.763Z Aletheia de DeepMind Resuelve 6 de 10 Problemas Matemáticos de Investigación y Rehúsa Falsificar los Demás https://aiexpert.news/en/article/anthropics-internal-ai-marketplace-found-smarter-agents-win-and-their-humans-don ai|expert en 2026-04-25T13:28:30.006Z Anthropic Study Finds AI Model Tier Drives Outcome Gaps Users Cannot Detect https://aiexpert.news/pt/article/anthropics-internal-ai-marketplace-found-smarter-agents-win-and-their-humans-don ai|expert pt-BR 2026-04-25T13:28:30.006Z Estudo da Anthropic Revela que Nível do Modelo de IA Determina Resultados que Usuários Não Conseguem Detectar https://aiexpert.news/es/article/anthropics-internal-ai-marketplace-found-smarter-agents-win-and-their-humans-don ai|expert es 2026-04-25T13:28:30.006Z Estudio de Anthropic Revela que el Nivel del Modelo de IA Determina Resultados que los Usuarios No Pueden Detectar https://aiexpert.news/en/article/deepseek-v4-launches-open-source-16t-param-pro-model-claims-parity-with-top-clos ai|expert en 2026-04-25T06:08:28.705Z DeepSeek V4-Pro Claims Benchmark Parity With Top Closed-Source Models on Math and STEM https://aiexpert.news/pt/article/deepseek-v4-launches-open-source-16t-param-pro-model-claims-parity-with-top-clos ai|expert pt-BR 2026-04-25T06:08:28.705Z DeepSeek V4-Pro Reivindica Paridade de Benchmark com os Melhores Modelos Fechados em Matemática e STEM https://aiexpert.news/es/article/deepseek-v4-launches-open-source-16t-param-pro-model-claims-parity-with-top-clos ai|expert es 2026-04-25T06:08:28.705Z DeepSeek V4-Pro Reivindica Paridad en Benchmarks con los Mejores Modelos Cerrados en Matemáticas y STEM https://aiexpert.news/en/article/ai-agents-are-closing-the-1-to-6-gap-that-made-saas-a-trillion-dollar-business ai|expert en 2026-04-25T05:58:28.702Z Sequoia's Julien Bek Targets the $6 in Services Behind Every $1 in SaaS https://aiexpert.news/pt/article/ai-agents-are-closing-the-1-to-6-gap-that-made-saas-a-trillion-dollar-business ai|expert pt-BR 2026-04-25T05:58:28.702Z Julien Bek, da Sequoia, mira os $6 em Serviços por trás de cada $1 em SaaS https://aiexpert.news/es/article/ai-agents-are-closing-the-1-to-6-gap-that-made-saas-a-trillion-dollar-business ai|expert es 2026-04-25T05:58:28.702Z Julien Bek, de Sequoia, apunta a los $6 en Servicios detrás de cada $1 en SaaS https://aiexpert.news/en/article/cohere-and-aleph-alpha-merge-in-20b-deal-to-build-a-transatlantic-rival-to-us-ai ai|expert en 2026-04-25T05:45:52.984Z Cohere and Aleph Alpha Merge in $20B Deal to Challenge U.S. AI Leaders https://aiexpert.news/pt/article/cohere-and-aleph-alpha-merge-in-20b-deal-to-build-a-transatlantic-rival-to-us-ai ai|expert pt-BR 2026-04-25T05:45:52.984Z Cohere e Aleph Alpha se fundem em negócio de $20B para desafiar líderes de IA dos EUA https://aiexpert.news/es/article/cohere-and-aleph-alpha-merge-in-20b-deal-to-build-a-transatlantic-rival-to-us-ai ai|expert es 2026-04-25T05:45:52.984Z Cohere y Aleph Alpha se fusionan en un acuerdo de $20B para desafiar a los líderes de IA de EE. UU. https://aiexpert.news/en/article/google-commits-up-to-40b-in-anthropic-competitor-landlord-and-now-biggest-backer ai|expert en 2026-04-25T05:05:18.708Z Google's $40 Billion Stake Makes It Anthropic's Investor, Chip Supplier, and Rival https://aiexpert.news/pt/article/google-commits-up-to-40b-in-anthropic-competitor-landlord-and-now-biggest-backer ai|expert pt-BR 2026-04-25T05:05:18.708Z A Participação de $40 Bilhões do Google Faz Dele Investidor, Fornecedor de Chips e Rival da Anthropic https://aiexpert.news/es/article/google-commits-up-to-40b-in-anthropic-competitor-landlord-and-now-biggest-backer ai|expert es 2026-04-25T05:05:18.708Z La Participación de $40 Mil Millones de Google lo Convierte en Inversor, Proveedor de Chips y Rival de Anthropic https://aiexpert.news/en/article/openais-gpt-55-deploys-codex-to-all-10000-nvidians-on-gb200-nvl72-at-35x-lower-t ai|expert en 2026-04-25T03:38:50.813Z GPT-5.5 Codex Reaches Every NVIDIA Employee at 35x Lower Token Cost https://aiexpert.news/pt/article/openais-gpt-55-deploys-codex-to-all-10000-nvidians-on-gb200-nvl72-at-35x-lower-t ai|expert pt-BR 2026-04-25T03:38:50.813Z GPT-5.5 Codex Chega a Todos os Funcionários da NVIDIA com Custo de Token 35x Menor https://aiexpert.news/es/article/openais-gpt-55-deploys-codex-to-all-10000-nvidians-on-gb200-nvl72-at-35x-lower-t ai|expert es 2026-04-25T03:38:50.813Z GPT-5.5 Codex Llega a Todos los Empleados de NVIDIA con un Costo de Token 35x Menor