https://aiexpert.news/en/article/elastic-attention-cores-slash-vision-transformer-costs-for-high-resolution-infer ai|expert en 2026-05-14T08:45:40.252Z VECA Cuts Vision Transformer Inference Cost to Linear Time https://aiexpert.news/pt/article/elastic-attention-cores-slash-vision-transformer-costs-for-high-resolution-infer ai|expert pt-BR 2026-05-14T08:45:40.252Z VECA Reduz Custo de Inferência em Vision Transformers para Tempo Linear https://aiexpert.news/es/article/elastic-attention-cores-slash-vision-transformer-costs-for-high-resolution-infer ai|expert es 2026-05-14T08:45:40.252Z VECA Reduce el Costo de Inferencia en Vision Transformers a Tiempo Lineal https://aiexpert.news/en/article/toolcua-solving-the-hybrid-action-space-problem-in-ai-computer-use-agents ai|expert en 2026-05-14T08:15:37.975Z ToolCUA reaches 46.85% on OSWorld, beats frontier agents on efficiency https://aiexpert.news/pt/article/toolcua-solving-the-hybrid-action-space-problem-in-ai-computer-use-agents ai|expert pt-BR 2026-05-14T08:15:37.975Z ToolCUA atinge 46.85% no OSWorld, supera agentes frontier em eficiência https://aiexpert.news/es/article/toolcua-solving-the-hybrid-action-space-problem-in-ai-computer-use-agents ai|expert es 2026-05-14T08:15:37.975Z ToolCUA alcanza 46.85% en OSWorld, supera agentes frontier en eficiencia https://aiexpert.news/en/article/meme-benchmark-evaluates-multi-entity-agent-memory-across-sessions ai|expert en 2026-05-14T07:40:38.518Z MEME benchmark finds 97% failure on agent memory dependency tasks https://aiexpert.news/pt/article/meme-benchmark-evaluates-multi-entity-agent-memory-across-sessions ai|expert pt-BR 2026-05-14T07:40:38.518Z Benchmark MEME encontra 97% de falha em tarefas de memória de agentes https://aiexpert.news/es/article/meme-benchmark-evaluates-multi-entity-agent-memory-across-sessions ai|expert es 2026-05-14T07:40:38.518Z Benchmark MEME encuentra 97% de fallos en tareas de memoria de agentes https://aiexpert.news/en/article/predicting-an-llms-post-training-potential-before-fine-tuning ai|expert en 2026-05-14T07:08:48.667Z RuDE Predicts Fine-Tuning Success Without Training https://aiexpert.news/pt/article/predicting-an-llms-post-training-potential-before-fine-tuning ai|expert pt-BR 2026-05-14T07:08:48.667Z RuDE Prevê Sucesso de Fine-Tuning Sem Treinamento https://aiexpert.news/es/article/predicting-an-llms-post-training-potential-before-fine-tuning ai|expert es 2026-05-14T07:08:48.667Z RuDE Predice Éxito de Fine-Tuning Sin Entrenamiento https://aiexpert.news/en/article/google-races-gemini-on-android-strategy-against-apples-ai-overhaul ai|expert en 2026-05-14T06:32:48.657Z Google Rebuilds Android Around Gemini Agent System https://aiexpert.news/pt/article/google-races-gemini-on-android-strategy-against-apples-ai-overhaul ai|expert pt-BR 2026-05-14T06:32:48.657Z Google Reconstrói Android em Torno do Sistema Gemini Agent https://aiexpert.news/es/article/google-races-gemini-on-android-strategy-against-apples-ai-overhaul ai|expert es 2026-05-14T06:32:48.657Z Google Reconstruye Android en Torno al Sistema Gemini Agent https://aiexpert.news/en/article/rubricem-trains-research-agents-on-evaluable-outputs-without-ground-truth ai|expert en 2026-05-14T05:45:39.100Z Google's RubricEM trains research agents without ground truth https://aiexpert.news/pt/article/rubricem-trains-research-agents-on-evaluable-outputs-without-ground-truth ai|expert pt-BR 2026-05-14T05:45:39.100Z RubricEM do Google treina agentes de pesquisa sem ground truth https://aiexpert.news/es/article/rubricem-trains-research-agents-on-evaluable-outputs-without-ground-truth ai|expert es 2026-05-14T05:45:39.100Z RubricEM de Google entrena agentes de investigación sin ground truth https://aiexpert.news/en/article/deepmind-spinout-isomorphic-labs-raises-21b-for-drug-discovery-ai ai|expert en 2026-05-14T05:15:36.846Z Isomorphic Labs Raises $2.1 Billion With No Drug Candidate Yet https://aiexpert.news/pt/article/deepmind-spinout-isomorphic-labs-raises-21b-for-drug-discovery-ai ai|expert pt-BR 2026-05-14T05:15:36.846Z Isomorphic Labs Levanta $2,1 Bilhão Sem Candidato a Fármaco Ainda https://aiexpert.news/es/article/deepmind-spinout-isomorphic-labs-raises-21b-for-drug-discovery-ai ai|expert es 2026-05-14T05:15:36.846Z Isomorphic Labs Recauda $2.100 Millones Sin Candidato a Fármaco Aún https://aiexpert.news/en/article/amazon-employees-admitting-to-ai-tool-overuse-to-inflate-internal-usage-scores ai|expert en 2026-05-14T04:44:48.596Z Amazon Developers Gaming AI Usage Metrics to Hit Internal Targets https://aiexpert.news/pt/article/amazon-employees-admitting-to-ai-tool-overuse-to-inflate-internal-usage-scores ai|expert pt-BR 2026-05-14T04:44:48.596Z Desenvolvedores da Amazon Inflando Métricas de Uso de IA para Atingir Metas Internas https://aiexpert.news/es/article/amazon-employees-admitting-to-ai-tool-overuse-to-inflate-internal-usage-scores ai|expert es 2026-05-14T04:44:48.596Z Desarrolladores de Amazon Inflando Métricas de Uso de IA para Cumplir Objetivos Internos https://aiexpert.news/en/article/nvidia-sap-team-on-trustworthy-specialized-agents ai|expert en 2026-05-14T04:12:48.557Z SAP and NVIDIA Close Enterprise Agent Governance Gap https://aiexpert.news/pt/article/nvidia-sap-team-on-trustworthy-specialized-agents ai|expert pt-BR 2026-05-14T04:12:48.557Z SAP e NVIDIA Fecham o Vácuo de Governança em Agentes Empresariais https://aiexpert.news/es/article/nvidia-sap-team-on-trustworthy-specialized-agents ai|expert es 2026-05-14T04:12:48.557Z SAP y NVIDIA Cierran la Brecha de Gobernanza en Agentes Empresariales https://aiexpert.news/en/article/european-ai-funding-surges-reshaping-startup-growth-despite-brain-drain ai|expert en 2026-05-14T03:40:48.454Z Fifty percent of European VC goes to AI in early 2026 https://aiexpert.news/pt/article/european-ai-funding-surges-reshaping-startup-growth-despite-brain-drain ai|expert pt-BR 2026-05-14T03:40:48.454Z Cinquenta por cento do capital de risco europeu vai para IA no início de 2026 https://aiexpert.news/es/article/european-ai-funding-surges-reshaping-startup-growth-despite-brain-drain ai|expert es 2026-05-14T03:40:48.454Z Cincuenta por ciento del capital de riesgo europeo va hacia IA a principios de 2026 https://aiexpert.news/en/article/llm-assisted-bug-hunting-collapses-vulnerability-disclosure-windows-to-30-days ai|expert en 2026-05-14T03:10:39.354Z LLM-assisted bug hunting makes 90-day disclosure window obsolete https://aiexpert.news/pt/article/llm-assisted-bug-hunting-collapses-vulnerability-disclosure-windows-to-30-days ai|expert pt-BR 2026-05-14T03:10:39.354Z LLM-assisted bug hunting torna a janela de divulgação de 90 dias obsoleta https://aiexpert.news/es/article/llm-assisted-bug-hunting-collapses-vulnerability-disclosure-windows-to-30-days ai|expert es 2026-05-14T03:10:39.354Z LLM-assisted bug hunting hace que la ventana de divulgación de 90 días sea obsoleta https://aiexpert.news/en/article/time-series-storage-design-choices-shape-cost-and-performance-for-ai-inference-p ai|expert en 2026-05-14T02:40:38.115Z Normalized schemas cut time-series storage 42 percent https://aiexpert.news/pt/article/time-series-storage-design-choices-shape-cost-and-performance-for-ai-inference-p ai|expert pt-BR 2026-05-14T02:40:38.115Z Esquemas normalizados reduzem armazenamento de time-series em 42 por cento https://aiexpert.news/es/article/time-series-storage-design-choices-shape-cost-and-performance-for-ai-inference-p ai|expert es 2026-05-14T02:40:38.115Z Esquemas normalizados reducen almacenamiento de series temporales 42 por ciento https://aiexpert.news/en/article/microsofts-1b-kenya-ai-data-center-stalls-over-power-shortage ai|expert en 2026-05-14T02:10:37.257Z Microsoft's $1 Billion Kenya Data Center Shelved Over Power Shortage https://aiexpert.news/pt/article/microsofts-1b-kenya-ai-data-center-stalls-over-power-shortage ai|expert pt-BR 2026-05-14T02:10:37.257Z Projeto de US$ 1 Bilhão de Data Center da Microsoft no Quênia é Suspenso por Falta de Energia https://aiexpert.news/es/article/microsofts-1b-kenya-ai-data-center-stalls-over-power-shortage ai|expert es 2026-05-14T02:10:37.257Z Proyecto de US$ 1 Mil Millones de Data Center de Microsoft en Kenia se Suspende por Falta de Energía https://aiexpert.news/en/article/beyond-red-teaming-formal-guarantees-for-llm-guardrails ai|expert en 2026-05-14T01:22:48.639Z Every Guardrail Classifier Tested Fails Formal Safety Verification https://aiexpert.news/pt/article/beyond-red-teaming-formal-guarantees-for-llm-guardrails ai|expert pt-BR 2026-05-14T01:22:48.639Z Cada Classificador de Guardrail Testado Falha em Verificação Formal de Segurança https://aiexpert.news/es/article/beyond-red-teaming-formal-guarantees-for-llm-guardrails ai|expert es 2026-05-14T01:22:48.639Z Cada Clasificador de Guardrail Probado Falla en Verificación Formal de Seguridad https://aiexpert.news/en/article/transformer-attention-concentration-predicts-token-behavior ai|expert en 2026-05-14T00:50:48.472Z Math Proof Shows Transformer Attention Stabilizes Predictably https://aiexpert.news/pt/article/transformer-attention-concentration-predicts-token-behavior ai|expert pt-BR 2026-05-14T00:50:48.472Z Prova Matemática Mostra que Atenção em Transformers Estabiliza Previsivelmente https://aiexpert.news/es/article/transformer-attention-concentration-predicts-token-behavior ai|expert es 2026-05-14T00:50:48.472Z Demostración Matemática Muestra que la Atención en Transformers se Estabiliza Predeciblemente https://aiexpert.news/en/article/engineering-rigor-in-ai-agents-columbia-researchers-warn-against-on-the-fly-synt ai|expert en 2026-05-14T00:20:39.375Z AI Agents Bypass Software Engineering, Risk Production Failure https://aiexpert.news/pt/article/engineering-rigor-in-ai-agents-columbia-researchers-warn-against-on-the-fly-synt ai|expert pt-BR 2026-05-14T00:20:39.375Z Agentes de IA Contornam Engenharia de Software, Arriscam Falhas em Produção https://aiexpert.news/es/article/engineering-rigor-in-ai-agents-columbia-researchers-warn-against-on-the-fly-synt ai|expert es 2026-05-14T00:20:39.375Z Agentes de IA Evaden la Ingeniería de Software, Arriesgan Fallos en Producción https://aiexpert.news/en/article/ai-chip-scarcity-in-2026-helium-rationing-660b-capex-walls-and-the-hbm-margin-tr ai|expert en 2026-05-13T19:20:38.477Z Chip Scarcity Hits Critical Point in 2026: $660B Capex Walls, Helium Cuts, 50% Slippage https://aiexpert.news/pt/article/ai-chip-scarcity-in-2026-helium-rationing-660b-capex-walls-and-the-hbm-margin-tr ai|expert pt-BR 2026-05-13T19:20:38.477Z Escassez de Chips Atinge Ponto Crítico em 2026: Muros de $660B em Capex, Cortes de Hélio, 50% de Atraso https://aiexpert.news/es/article/ai-chip-scarcity-in-2026-helium-rationing-660b-capex-walls-and-the-hbm-margin-tr ai|expert es 2026-05-13T19:20:38.477Z Escasez de Chips Llega a Punto Crítico en 2026: Muros de $660B en Capex, Cortes de Helio, 50% de Retraso https://aiexpert.news/en/article/european-ai-startups-aggressively-poach-talent-from-big-tech ai|expert en 2026-05-13T07:05:37.699Z Giannandrea's Move Signals European AI Startups' Silicon Valley Recruitment Push https://aiexpert.news/pt/article/european-ai-startups-aggressively-poach-talent-from-big-tech ai|expert pt-BR 2026-05-13T07:05:37.699Z O movimento de Giannandrea sinaliza push de recrutamento no Vale do Silício de startups europeias de IA https://aiexpert.news/es/article/european-ai-startups-aggressively-poach-talent-from-big-tech ai|expert es 2026-05-13T07:05:37.699Z El movimiento de Giannandrea señala el impulso de reclutamiento en Silicon Valley de startups europeas de IA https://aiexpert.news/en/article/dynamic-skill-management-for-llm-agents-improves-transparency ai|expert en 2026-05-13T06:34:48.548Z SLIM improves LLM agent performance 7 percentage points https://aiexpert.news/pt/article/dynamic-skill-management-for-llm-agents-improves-transparency ai|expert pt-BR 2026-05-13T06:34:48.548Z SLIM melhora performance de agentes LLM em 7 pontos percentuais https://aiexpert.news/es/article/dynamic-skill-management-for-llm-agents-improves-transparency ai|expert es 2026-05-13T06:34:48.548Z SLIM mejora desempeño de agentes LLM en 7 puntos porcentuales https://aiexpert.news/en/article/shepherd-formalizes-ai-agent-execution-traces-for-auditability ai|expert en 2026-05-13T06:02:48.586Z Shepherd Raises Agent Accuracy 90% With Forking Traces https://aiexpert.news/pt/article/shepherd-formalizes-ai-agent-execution-traces-for-auditability ai|expert pt-BR 2026-05-13T06:02:48.586Z Shepherd Aumenta Precisão de Agentes em 90% Com Rastreamento por Bifurcação https://aiexpert.news/es/article/shepherd-formalizes-ai-agent-execution-traces-for-auditability ai|expert es 2026-05-13T06:02:48.586Z Shepherd Aumenta Precisión de Agentes 90% Con Rastreo Por Ramificación https://aiexpert.news/en/article/wildclawbench-first-real-world-benchmark-for-long-horizon-ai-agents ai|expert en 2026-05-13T05:30:48.535Z WildClawBench: Claude Opus Clears 62% in Real-World Agent Evaluation https://aiexpert.news/pt/article/wildclawbench-first-real-world-benchmark-for-long-horizon-ai-agents ai|expert pt-BR 2026-05-13T05:30:48.535Z WildClawBench: Claude Opus Atinge 62% em Avaliação de Agent em Ambiente Real https://aiexpert.news/es/article/wildclawbench-first-real-world-benchmark-for-long-horizon-ai-agents ai|expert es 2026-05-13T05:30:48.535Z WildClawBench: Claude Opus Alcanza 62% en Evaluación de Agentes en Mundo Real https://aiexpert.news/en/article/deco-sparse-moe-models-match-dense-transformer-speed-on-edge-devices ai|expert en 2026-05-13T05:00:37.714Z Sparse MoE Models Match Dense Transformers at 3× Faster Inference https://aiexpert.news/pt/article/deco-sparse-moe-models-match-dense-transformer-speed-on-edge-devices ai|expert pt-BR 2026-05-13T05:00:37.714Z Modelos Sparse MoE Emparelham com Transformers Densos a 3× Mais Rápido em Inferência https://aiexpert.news/es/article/deco-sparse-moe-models-match-dense-transformer-speed-on-edge-devices ai|expert es 2026-05-13T05:00:37.714Z Modelos Sparse MoE Equiparan Desempeño con Transformers Densos a 3× Más Rápido en Inferencia https://aiexpert.news/en/article/beyond-adamw-survey-of-memory-efficient-and-curvature-based-llm-optimizers ai|expert en 2026-05-13T04:28:48.576Z Muon Optimizer Achieves 2× Speed Over AdamW in Production LLM Training https://aiexpert.news/pt/article/beyond-adamw-survey-of-memory-efficient-and-curvature-based-llm-optimizers ai|expert pt-BR 2026-05-13T04:28:48.576Z Otimizador Muon Atinge 2× de Velocidade sobre AdamW no Treinamento de LLM em Produção https://aiexpert.news/es/article/beyond-adamw-survey-of-memory-efficient-and-curvature-based-llm-optimizers ai|expert es 2026-05-13T04:28:48.576Z Optimizador Muon Logra 2× de Velocidad sobre AdamW en Entrenamiento de LLM en Producción https://aiexpert.news/en/article/civex-causal-verification-for-tool-using-language-agents ai|expert en 2026-05-13T03:50:48.372Z CIVeX Logs Zero False Executions in Confounded Workflows https://aiexpert.news/pt/article/civex-causal-verification-for-tool-using-language-agents ai|expert pt-BR 2026-05-13T03:50:48.372Z CIVeX Registra Zero Execuções Falsas em Workflows Confundidos https://aiexpert.news/es/article/civex-causal-verification-for-tool-using-language-agents ai|expert es 2026-05-13T03:50:48.372Z CIVeX Registra Cero Ejecuciones Falsas en Flujos Confundidos https://aiexpert.news/en/article/jensen-huang-joins-trump-china-trip-nvidia-signals-trade-policy-influence ai|expert en 2026-05-13T03:20:37.411Z Huang joins Trump's Beijing summit as Nvidia seeks China clarity https://aiexpert.news/pt/article/jensen-huang-joins-trump-china-trip-nvidia-signals-trade-policy-influence ai|expert pt-BR 2026-05-13T03:20:37.411Z Huang se junta à delegação chinesa de Trump enquanto Nvidia busca clareza sobre mercado chinês https://aiexpert.news/es/article/jensen-huang-joins-trump-china-trip-nvidia-signals-trade-policy-influence ai|expert es 2026-05-13T03:20:37.411Z Huang se une a la delegación china de Trump mientras Nvidia busca claridad sobre el mercado chino https://aiexpert.news/en/article/prediction-bottlenecks-dont-discover-causal-structurenew-verification-needed ai|expert en 2026-05-13T03:04:49.120Z Paper Dismantles Causal Discovery Claim in Prediction Models https://aiexpert.news/pt/article/prediction-bottlenecks-dont-discover-causal-structurenew-verification-needed ai|expert pt-BR 2026-05-13T03:04:49.120Z Artigo Desmonta Alegação sobre Descoberta Causal em Modelos de Previsão https://aiexpert.news/es/article/prediction-bottlenecks-dont-discover-causal-structurenew-verification-needed ai|expert es 2026-05-13T03:04:49.120Z Artículo Desmantela Afirmación sobre Descubrimiento Causal en Modelos de Predicción https://aiexpert.news/en/article/emergent-semantic-role-understanding-appears-without-task-specific-fine-tuning ai|expert en 2026-05-13T02:32:48.422Z Frozen Models Encode Semantic Roles Without Fine-Tuning https://aiexpert.news/pt/article/emergent-semantic-role-understanding-appears-without-task-specific-fine-tuning ai|expert pt-BR 2026-05-13T02:32:48.422Z Modelos Congelados Codificam Papéis Semânticos Sem Fine-Tuning https://aiexpert.news/es/article/emergent-semantic-role-understanding-appears-without-task-specific-fine-tuning ai|expert es 2026-05-13T02:32:48.422Z Modelos Congelados Codifican Roles Semánticos Sin Fine-Tuning https://aiexpert.news/en/article/google-disrupts-hacker-use-of-ai-for-scaled-mass-exploitation-campaigns ai|expert en 2026-05-13T01:50:37.009Z Google Confirms First AI-Written Zero-Day Exploit https://aiexpert.news/pt/article/google-disrupts-hacker-use-of-ai-for-scaled-mass-exploitation-campaigns ai|expert pt-BR 2026-05-13T01:50:37.009Z Google Confirma Primeiro Zero-Day Escrito por IA https://aiexpert.news/es/article/google-disrupts-hacker-use-of-ai-for-scaled-mass-exploitation-campaigns ai|expert es 2026-05-13T01:50:37.009Z Google Confirma Primer Zero-Day Escrito por IA https://aiexpert.news/en/article/arm-cpus-capture-2b-in-ai-data-center-sales-yet-lag-95-market-share ai|expert en 2026-05-13T00:50:48.561Z Arm AGI CPU Lands $2B in Demand, Claims 4% Server Market Share https://aiexpert.news/pt/article/arm-cpus-capture-2b-in-ai-data-center-sales-yet-lag-95-market-share ai|expert pt-BR 2026-05-13T00:50:48.561Z CPU Arm AGI Atrai $2B em Demanda, Conquista 4% da Fatia de Mercado em Servidores https://aiexpert.news/es/article/arm-cpus-capture-2b-in-ai-data-center-sales-yet-lag-95-market-share ai|expert es 2026-05-13T00:50:48.561Z CPU Arm AGI Atrae $2B en Demanda, Captura 4% de Cuota de Mercado en Servidores https://aiexpert.news/en/article/mistral-tanstack-supply-chain-attack-exposes-cicd-credentials ai|expert en 2026-05-13T00:20:38.328Z Mini Shai-Hulud supply-chain attack targets Mistral and TanStack https://aiexpert.news/pt/article/mistral-tanstack-supply-chain-attack-exposes-cicd-credentials ai|expert pt-BR 2026-05-13T00:20:38.328Z Ataque à cadeia de suprimentos Mini Shai-Hulud atinge Mistral e TanStack https://aiexpert.news/es/article/mistral-tanstack-supply-chain-attack-exposes-cicd-credentials ai|expert es 2026-05-13T00:20:38.328Z Ataque a la cadeia de suministro Mini Shai-Hulud apunta a Mistral y TanStack https://aiexpert.news/en/article/local-first-ai-inference-cost-efficient-document-processing-at-the-edge ai|expert en 2026-05-13T00:00:37.865Z Hybrid AI Routing Cuts Document Processing Costs 75% https://aiexpert.news/pt/article/local-first-ai-inference-cost-efficient-document-processing-at-the-edge ai|expert pt-BR 2026-05-13T00:00:37.865Z Roteamento Híbrido de IA Reduz Custos 75% em Processamento de Documentos https://aiexpert.news/es/article/local-first-ai-inference-cost-efficient-document-processing-at-the-edge ai|expert es 2026-05-13T00:00:37.865Z Enrutamiento Híbrido de IA Reduce Costos 75% en Procesamiento de Documentos