https://aiexpert.news/en/article/open-weights-models-hit-new-deployment-milestones-summer-2026-landscape-report ai|expert en 2026-08-14T16:28:37.645Z Production Use, Not Hype, Drives Open-Model Adoption https://aiexpert.news/pt/article/open-weights-models-hit-new-deployment-milestones-summer-2026-landscape-report ai|expert pt-BR 2026-08-14T16:28:37.645Z Uso em Produção, Não Hype, Impulsiona a Adoção de Modelos Abertos https://aiexpert.news/es/article/open-weights-models-hit-new-deployment-milestones-summer-2026-landscape-report ai|expert es 2026-08-14T16:28:37.645Z El Uso en Producción, No la Expectativa, Impulsa la Adopción de Modelos Abiertos https://aiexpert.news/en/article/from-tables-to-questions-scottish-waters-conversational-data-analytics ai|expert en 2026-08-14T15:46:37.838Z SPARK Cuts Data Queries from 8 Clicks to Single Question https://aiexpert.news/pt/article/from-tables-to-questions-scottish-waters-conversational-data-analytics ai|expert pt-BR 2026-08-14T15:46:37.838Z SPARK Reduz Consultas de Dados de 8 Cliques para uma Pergunta https://aiexpert.news/es/article/from-tables-to-questions-scottish-waters-conversational-data-analytics ai|expert es 2026-08-14T15:46:37.838Z SPARK Reduce Consultas de Datos de 8 Clics a una Sola Pregunta https://aiexpert.news/en/article/quotebench-the-escaping-gap-in-llm-coding-agents ai|expert en 2026-08-14T15:15:23.383Z QuoteBench Exposes 55-Point Gap Hidden in Coding-Agent Benchmarks https://aiexpert.news/pt/article/quotebench-the-escaping-gap-in-llm-coding-agents ai|expert pt-BR 2026-08-14T15:15:23.383Z QuoteBench Expõe Brecha de 55 Pontos Escondida em Benchmarks de Agentes de Código https://aiexpert.news/es/article/quotebench-the-escaping-gap-in-llm-coding-agents ai|expert es 2026-08-14T15:15:23.383Z QuoteBench Expone una Brecha de 55 Puntos Oculta en Benchmarks de Agentes de Código https://aiexpert.news/en/article/google-deepminds-sign-language-ai-system-now-live-in-production ai|expert en 2026-08-14T15:02:37.467Z Google DeepMind Ships Sign Language AI to Pixel 11 Users https://aiexpert.news/pt/article/google-deepminds-sign-language-ai-system-now-live-in-production ai|expert pt-BR 2026-08-14T15:02:37.467Z Google DeepMind lança IA de linguagem de sinais no Pixel 11 https://aiexpert.news/es/article/google-deepminds-sign-language-ai-system-now-live-in-production ai|expert es 2026-08-14T15:02:37.467Z Google DeepMind lanza IA de lenguaje de signos en Pixel 11 https://aiexpert.news/en/article/how-cloudflare-secures-model-context-protocol-traffic-at-scale ai|expert en 2026-08-14T14:31:29.747Z Cloudflare Maps Three-Layer Defense Against Unauthorized MCP Servers https://aiexpert.news/pt/article/how-cloudflare-secures-model-context-protocol-traffic-at-scale ai|expert pt-BR 2026-08-14T14:31:29.747Z Cloudflare Mapeia Defesa em Três Camadas Contra Servidores MCP Não Autorizados https://aiexpert.news/es/article/how-cloudflare-secures-model-context-protocol-traffic-at-scale ai|expert es 2026-08-14T14:31:29.747Z Cloudflare Mapea Defensa en Tres Capas Contra Servidores MCP No Autorizados https://aiexpert.news/en/article/asml-amadeus-back-mistral-data-center-capex ai|expert en 2026-08-14T14:00:37.595Z ASML, Amadeus Back Mistral's $50B European Compute Build https://aiexpert.news/pt/article/asml-amadeus-back-mistral-data-center-capex ai|expert pt-BR 2026-08-14T14:00:37.595Z ASML, Amadeus Apostam em Construção de $50B de Computação Europeia da Mistral https://aiexpert.news/es/article/asml-amadeus-back-mistral-data-center-capex ai|expert es 2026-08-14T14:00:37.595Z ASML, Amadeus Respaldan la Construcción de $50B de Cómputo Europeo de Mistral https://aiexpert.news/en/article/agents-now-ship-code-proofs-not-just-functions ai|expert en 2026-08-14T13:30:24.669Z Agents Solve Only 27 of 43 Verified Code Tasks https://aiexpert.news/pt/article/agents-now-ship-code-proofs-not-just-functions ai|expert pt-BR 2026-08-14T13:30:24.669Z Agentes Resolvem Apenas 27 de 43 Tarefas de Código Verificado https://aiexpert.news/es/article/agents-now-ship-code-proofs-not-just-functions ai|expert es 2026-08-14T13:30:24.669Z Agentes Resuelven Solo 27 de 43 Tareas de Código Verificado https://aiexpert.news/en/article/near-packaged-optics-gains-ground-as-ai-chip-infra-shifts ai|expert en 2026-08-14T12:44:37.539Z NPO Gains Traction as Yield Challenges Dent CPO Economics https://aiexpert.news/pt/article/near-packaged-optics-gains-ground-as-ai-chip-infra-shifts ai|expert pt-BR 2026-08-14T12:44:37.539Z NPO Ganha Tração com Desafios de Rendimento Abalando a Economia de CPO https://aiexpert.news/es/article/near-packaged-optics-gains-ground-as-ai-chip-infra-shifts ai|expert es 2026-08-14T12:44:37.539Z NPO Gana Tracción mientras Desafíos de Rendimiento Debilitan la Economía de CPO https://aiexpert.news/en/article/smart-routing-cost-optimized-model-selection-at-runtime ai|expert en 2026-08-14T12:12:39.235Z Databricks Smart Routing cuts coding agent costs 56% on public benchmarks https://aiexpert.news/pt/article/smart-routing-cost-optimized-model-selection-at-runtime ai|expert pt-BR 2026-08-14T12:12:39.235Z Roteamento Inteligente do Databricks reduz custos de agentes de codificação em 56% em benchmarks públicos https://aiexpert.news/es/article/smart-routing-cost-optimized-model-selection-at-runtime ai|expert es 2026-08-14T12:12:39.235Z El Enrutamiento Inteligente de Databricks reduce los costos de agentes de codificación en 56% en puntos de referencia públicos https://aiexpert.news/en/article/strands-lerobot-end-to-end-robot-learning-stack-from-data-to-deployment ai|expert en 2026-08-13T18:58:37.595Z AWS and Hugging Face launch Strands, end-to-end robot learning SDK https://aiexpert.news/pt/article/strands-lerobot-end-to-end-robot-learning-stack-from-data-to-deployment ai|expert pt-BR 2026-08-13T18:58:37.595Z AWS e Hugging Face lançam Strands, SDK de aprendizado robótico end-to-end https://aiexpert.news/es/article/strands-lerobot-end-to-end-robot-learning-stack-from-data-to-deployment ai|expert es 2026-08-13T18:58:37.595Z AWS y Hugging Face lanzan Strands, SDK de aprendizaje robótico end-to-end https://aiexpert.news/en/article/reproducing-2200-icml-papers-hugging-face-audit-of-ml-reproducibility ai|expert en 2026-08-13T18:26:37.875Z Hugging Face Audits 2,226 ICML Papers, Finds 23% With False Claims https://aiexpert.news/pt/article/reproducing-2200-icml-papers-hugging-face-audit-of-ml-reproducibility ai|expert pt-BR 2026-08-13T18:26:37.875Z Hugging Face Audita 2.226 Artigos do ICML, Encontra 23% Com Afirmações Falsas https://aiexpert.news/es/article/reproducing-2200-icml-papers-hugging-face-audit-of-ml-reproducibility ai|expert es 2026-08-13T18:26:37.875Z Hugging Face Audita 2.226 Artículos de ICML, Encuentra 23% Con Afirmaciones Falsas https://aiexpert.news/en/article/structural-silence-ai-infrastructure-disadvantages-underrepresented-languages-at ai|expert en 2026-08-13T13:52:37.543Z Training Data Bias Compounds at Every Layer for Underrepresented Languages https://aiexpert.news/pt/article/structural-silence-ai-infrastructure-disadvantages-underrepresented-languages-at ai|expert pt-BR 2026-08-13T13:52:37.543Z Viés em Dados de Treinamento se Agrava em Cada Camada para Linguagens Subrepresentadas https://aiexpert.news/es/article/structural-silence-ai-infrastructure-disadvantages-underrepresented-languages-at ai|expert es 2026-08-13T13:52:37.543Z El Sesgo en Datos de Entrenamiento se Agrava en Cada Capa para Lenguajes Subrepresentados https://aiexpert.news/en/article/how-ai-disrupts-engineering-progression-rethinking-career-ladders-in-an-agent-na ai|expert en 2026-08-13T13:20:37.964Z AI Is Collapsing Engineering Career Ladders https://aiexpert.news/pt/article/how-ai-disrupts-engineering-progression-rethinking-career-ladders-in-an-agent-na ai|expert pt-BR 2026-08-13T13:20:37.964Z IA Está Desmoronando as Escadas de Carreira em Engenharia https://aiexpert.news/es/article/how-ai-disrupts-engineering-progression-rethinking-career-ladders-in-an-agent-na ai|expert es 2026-08-13T13:20:37.964Z La IA Está Colapsando las Escaleras de Carrera en Ingeniería https://aiexpert.news/en/article/claudes-sandbox-breach-containment-limits-under-eval ai|expert en 2026-08-13T10:48:37.585Z Claude Reached Production Systems for Three Months Before Detection https://aiexpert.news/pt/article/claudes-sandbox-breach-containment-limits-under-eval ai|expert pt-BR 2026-08-13T10:48:37.585Z Claude Alcançou Sistemas em Produção por Três Meses Antes da Detecção https://aiexpert.news/es/article/claudes-sandbox-breach-containment-limits-under-eval ai|expert es 2026-08-13T10:48:37.585Z Claude Alcanzó Sistemas en Producción Durante Tres Meses Antes de ser Detectado https://aiexpert.news/en/article/convergent-detour-hijacking-supply-chain-risk-in-llm-agent-tool-ecosystems ai|expert en 2026-08-13T07:26:37.527Z LLM Agents Fall to Supply-Chain Attacks Hiding in Plain Language https://aiexpert.news/pt/article/convergent-detour-hijacking-supply-chain-risk-in-llm-agent-tool-ecosystems ai|expert pt-BR 2026-08-13T07:26:37.527Z Agentes LLM Caem para Supply-Chain Attacks Escondidos em Linguagem Simples https://aiexpert.news/es/article/convergent-detour-hijacking-supply-chain-risk-in-llm-agent-tool-ecosystems ai|expert es 2026-08-13T07:26:37.527Z Agentes LLM Sucumben a Supply-Chain Attacks Ocultos en Lenguaje Plano https://aiexpert.news/en/article/test-time-model-scaffolding-strong-to-weak-reasoning-without-retraining ai|expert en 2026-08-13T05:26:37.602Z Inference-Time Scaffolding Lifts Weak Model Accuracy to 0.91 https://aiexpert.news/pt/article/test-time-model-scaffolding-strong-to-weak-reasoning-without-retraining ai|expert pt-BR 2026-08-13T05:26:37.602Z Andaimes no Tempo de Inferência Elevam Acurácia de Modelos Fracos a 0.91 https://aiexpert.news/es/article/test-time-model-scaffolding-strong-to-weak-reasoning-without-retraining ai|expert es 2026-08-13T05:26:37.602Z Andamios en Tiempo de Inferencia Elevan la Precisión de Modelos Débiles a 0.91 https://aiexpert.news/en/article/vakra-benchmark-tests-agent-tool-use-under-real-constraints-8000-apis-62-domains ai|expert en 2026-08-13T03:48:37.497Z IBM's VAKRA Benchmark Shows Models Fail 97% of Policy-Constrained Tasks https://aiexpert.news/pt/article/vakra-benchmark-tests-agent-tool-use-under-real-constraints-8000-apis-62-domains ai|expert pt-BR 2026-08-13T03:48:37.497Z O Benchmark VAKRA da IBM Mostra que Modelos Falham em 97% das Tarefas com Restrições de Política https://aiexpert.news/es/article/vakra-benchmark-tests-agent-tool-use-under-real-constraints-8000-apis-62-domains ai|expert es 2026-08-13T03:48:37.497Z El Benchmark VAKRA de IBM Demuestra que los Modelos Fallan en el 97% de las Tarefas Restringidas por Políticas https://aiexpert.news/en/article/redistribution-based-cost-inference-enables-safe-rl-training-with-sparse-feedbac ai|expert en 2026-08-13T03:16:38.961Z RCI Framework Cuts Constraint Violations in Offline Safe RL Training https://aiexpert.news/pt/article/redistribution-based-cost-inference-enables-safe-rl-training-with-sparse-feedbac ai|expert pt-BR 2026-08-13T03:16:38.961Z O Framework RCI Reduz Violações de Restrições no Treinamento de RL Offline Seguro https://aiexpert.news/es/article/redistribution-based-cost-inference-enables-safe-rl-training-with-sparse-feedbac ai|expert es 2026-08-13T03:16:38.961Z El Framework RCI Reduce Violaciones de Restricciones en Entrenamientos de RL Seguro Offline https://aiexpert.news/en/article/mcp-goes-stateless-developers-question-whether-its-still-a-protocol ai|expert en 2026-08-13T02:16:38.814Z MCP 2.0 Ditches Sessions for Serverless Scaling https://aiexpert.news/pt/article/mcp-goes-stateless-developers-question-whether-its-still-a-protocol ai|expert pt-BR 2026-08-13T02:16:38.814Z MCP 2.0 Abandona Sessões para Escala Serverless https://aiexpert.news/es/article/mcp-goes-stateless-developers-question-whether-its-still-a-protocol ai|expert es 2026-08-13T02:16:38.814Z MCP 2.0 Abandona Sesiones para Escalado Serverless https://aiexpert.news/en/article/lfm25-vl-3b-edge-vision-model-optimized-for-resource-constrained-deployment ai|expert en 2026-08-13T01:46:37.878Z Liquid AI ships 3.1B edge vision model with 228-token/second throughput https://aiexpert.news/pt/article/lfm25-vl-3b-edge-vision-model-optimized-for-resource-constrained-deployment ai|expert pt-BR 2026-08-13T01:46:37.878Z Liquid AI lança modelo de visão de borda com 3.1B parâmetros e 228 tokens/segundo https://aiexpert.news/es/article/lfm25-vl-3b-edge-vision-model-optimized-for-resource-constrained-deployment ai|expert es 2026-08-13T01:46:37.878Z Liquid AI lanza modelo de visión edge de 3.1B parámetros con throughput de 228 tokens/segundo https://aiexpert.news/en/article/olmoearth-embeddings-fine-tunable-geospatial-representations-for-satellite-analy ai|expert en 2026-08-12T23:38:37.378Z Open Geospatial Embeddings Lower Barrier to Satellite Data Analysis https://aiexpert.news/pt/article/olmoearth-embeddings-fine-tunable-geospatial-representations-for-satellite-analy ai|expert pt-BR 2026-08-12T23:38:37.378Z Embeddings Geoespaciais Abertos Reduzem Barreiras para Análise de Dados de Satélite https://aiexpert.news/es/article/olmoearth-embeddings-fine-tunable-geospatial-representations-for-satellite-analy ai|expert es 2026-08-12T23:38:37.378Z Los Embeddings Geoespaciales Abiertos Reducen Barreras para el Análisis de Datos de Satélite https://aiexpert.news/en/article/how-ai-agents-tackled-a-40-year-math-problem-grothendieck-constant-case-study ai|expert en 2026-08-12T23:06:37.550Z AI System Narrows Bounds on 40-Year Math Problem https://aiexpert.news/pt/article/how-ai-agents-tackled-a-40-year-math-problem-grothendieck-constant-case-study ai|expert pt-BR 2026-08-12T23:06:37.550Z Sistema de IA Reduz Limites de Problema Matemático de 40 Anos https://aiexpert.news/es/article/how-ai-agents-tackled-a-40-year-math-problem-grothendieck-constant-case-study ai|expert es 2026-08-12T23:06:37.550Z Sistema de IA Reduce Límites en Problema Matemático de 40 Años https://aiexpert.news/en/article/meta-scam-detection-at-billion-scale-on-whatsapp-with-e2e-encryption ai|expert en 2026-08-12T22:34:37.463Z Meta Ships On-Device Scam Detection for WhatsApp's 3 Billion Users https://aiexpert.news/pt/article/meta-scam-detection-at-billion-scale-on-whatsapp-with-e2e-encryption ai|expert pt-BR 2026-08-12T22:34:37.463Z Meta Lança Detecção de Golpes em Dispositivo para os 3 Bilhões de Usuários do WhatsApp https://aiexpert.news/es/article/meta-scam-detection-at-billion-scale-on-whatsapp-with-e2e-encryption ai|expert es 2026-08-12T22:34:37.463Z Meta Lanza Detección de Estafas en Dispositivo para los 3 Mil Millones de Usuarios de WhatsApp https://aiexpert.news/en/article/spotifys-external-index-low-latency-data-lake-queries-without-duplication ai|expert en 2026-08-12T22:02:38.231Z Spotify Cuts Data Lake Query Latency to Sub-Second Without Duplication https://aiexpert.news/pt/article/spotifys-external-index-low-latency-data-lake-queries-without-duplication ai|expert pt-BR 2026-08-12T22:02:38.231Z Spotify Reduz Latência de Consultas em Data Lake para Sub-segundo Sem Duplicação https://aiexpert.news/es/article/spotifys-external-index-low-latency-data-lake-queries-without-duplication ai|expert es 2026-08-12T22:02:38.231Z Spotify Reduce la Latencia de Consultas en Data Lake a Sub-segundo Sin Duplicación https://aiexpert.news/en/article/amtraks-50-year-data-backbone-how-legacy-systems-drive-modern-architecture ai|expert en 2026-08-12T21:30:37.401Z Amtrak Consolidates Rail Data Into Single Databricks Platform https://aiexpert.news/pt/article/amtraks-50-year-data-backbone-how-legacy-systems-drive-modern-architecture ai|expert pt-BR 2026-08-12T21:30:37.401Z Amtrak Consolida Dados Ferroviários em Plataforma Databricks Única https://aiexpert.news/es/article/amtraks-50-year-data-backbone-how-legacy-systems-drive-modern-architecture ai|expert es 2026-08-12T21:30:37.401Z Amtrak Consolida Datos Ferroviarios en una Plataforma Databricks Única https://aiexpert.news/en/article/databricks-scales-network-config-to-millions-of-vmslessons-for-ml-ops ai|expert en 2026-08-12T21:00:24.510Z Databricks cuts network-config latency 98.5% with async snapshots https://aiexpert.news/pt/article/databricks-scales-network-config-to-millions-of-vmslessons-for-ml-ops ai|expert pt-BR 2026-08-12T21:00:24.510Z Databricks reduz latência de configuração de rede em 98,5% com snapshots assíncronos https://aiexpert.news/es/article/databricks-scales-network-config-to-millions-of-vmslessons-for-ml-ops ai|expert es 2026-08-12T21:00:24.510Z Databricks reduce la latencia de configuración de red en un 98,5% con snapshots asíncronos https://aiexpert.news/en/article/meta-cuts-server-count-25-through-memory-reuseheres-the-pattern ai|expert en 2026-08-12T20:28:37.547Z Meta reduces servers 25% with CXL memory expansion https://aiexpert.news/pt/article/meta-cuts-server-count-25-through-memory-reuseheres-the-pattern ai|expert pt-BR 2026-08-12T20:28:37.547Z Meta reduz servidores em 25% com expansão de memória CXL https://aiexpert.news/es/article/meta-cuts-server-count-25-through-memory-reuseheres-the-pattern ai|expert es 2026-08-12T20:28:37.547Z Meta reduce servidores en 25% con expansión de memoria CXL