https://aiexpert.news/en/article/agents-k1-orchestrating-scientific-knowledge-for-agent-systems ai|expert en 2026-06-12T13:50:10.671Z Agents-K1 Replaces RAG Text Chunks With Typed Scientific Knowledge Graphs https://aiexpert.news/pt/article/agents-k1-orchestrating-scientific-knowledge-for-agent-systems ai|expert pt-BR 2026-06-12T13:50:10.671Z Agents-K1 Substitui Fatias de Texto RAG por Grafos de Conhecimento Científico Digitados https://aiexpert.news/es/article/agents-k1-orchestrating-scientific-knowledge-for-agent-systems ai|expert es 2026-06-12T13:50:10.671Z Agents-K1 Reemplaza Fragmentos de Texto RAG con Grafos de Conocimiento Científico Tipados https://aiexpert.news/en/article/pinecone-onelake-embedding-agents-in-enterprise-data-pipelines ai|expert en 2026-06-12T12:58:10.497Z Pinecone and Microsoft Claim 95% Token Reduction for LLM Workloads https://aiexpert.news/pt/article/pinecone-onelake-embedding-agents-in-enterprise-data-pipelines ai|expert pt-BR 2026-06-12T12:58:10.497Z Pinecone e Microsoft Afirmam Redução de 95% em Tokens para Cargas de Trabalho de LLM https://aiexpert.news/es/article/pinecone-onelake-embedding-agents-in-enterprise-data-pipelines ai|expert es 2026-06-12T12:58:10.497Z Pinecone y Microsoft afirman una reducción del 95% de tokens para cargas de trabajo de LLM https://aiexpert.news/en/article/agentic-refactoring-modernizing-legacy-code-in-weeks-not-years ai|expert en 2026-06-12T11:28:10.410Z ServiceTitan Cuts Legacy Migration from Years to Weeks With AI https://aiexpert.news/pt/article/agentic-refactoring-modernizing-legacy-code-in-weeks-not-years ai|expert pt-BR 2026-06-12T11:28:10.410Z ServiceTitan Reduz Migração de Legado de Anos para Semanas com IA https://aiexpert.news/es/article/agentic-refactoring-modernizing-legacy-code-in-weeks-not-years ai|expert es 2026-06-12T11:28:10.410Z ServiceTitan reduce la migración de legado de años a semanas con IA https://aiexpert.news/en/article/evoarena-benchmarks-llm-agent-robustness-in-dynamic-environments ai|expert en 2026-06-12T06:18:10.526Z EvoArena Benchmark Exposes Agent Collapse in Evolving Environments https://aiexpert.news/pt/article/evoarena-benchmarks-llm-agent-robustness-in-dynamic-environments ai|expert pt-BR 2026-06-12T06:18:10.526Z EvoArena Benchmark Revela Colapso de Agentes em Ambientes em Evolução https://aiexpert.news/es/article/evoarena-benchmarks-llm-agent-robustness-in-dynamic-environments ai|expert es 2026-06-12T06:18:10.526Z EvoArena Benchmark Revela Colapso de Agentes en Entornos en Evolución https://aiexpert.news/en/article/hypertool-mcp-style-orchestration-for-efficient-agent-tool-traces ai|expert en 2026-06-12T05:28:10.522Z HyperTool Doubles Qwen Accuracy by Bundling Tool Calls https://aiexpert.news/pt/article/hypertool-mcp-style-orchestration-for-efficient-agent-tool-traces ai|expert pt-BR 2026-06-12T05:28:10.522Z HyperTool Dobro o Acerto do Qwen ao Agrupar Chamadas de Ferramentas https://aiexpert.news/es/article/hypertool-mcp-style-orchestration-for-efficient-agent-tool-traces ai|expert es 2026-06-12T05:28:10.522Z HyperTool duplica la precisión de Qwen al unificar llamadas a herramientas https://aiexpert.news/en/article/agent-environment-engineering-bottleneck-shifts-from-capabilities-to-execution-d ai|expert en 2026-06-12T04:12:10.313Z Sub-$11 Agent Outperforms Specialized Research Frameworks https://aiexpert.news/pt/article/agent-environment-engineering-bottleneck-shifts-from-capabilities-to-execution-d ai|expert pt-BR 2026-06-12T04:12:10.313Z Agente Sub-$11 Supera Quadros de Pesquisa Especializados https://aiexpert.news/es/article/agent-environment-engineering-bottleneck-shifts-from-capabilities-to-execution-d ai|expert es 2026-06-12T04:12:10.313Z Sub-$11 Agente Supera Marcos de Investigación Especializados https://aiexpert.news/en/article/recursive-agent-harnesses-multi-level-agent-decomposition-at-production-scale ai|expert en 2026-06-12T03:40:10.476Z Recursive Agent Harness Achieves 89% Accuracy on Long-Context Code Tasks https://aiexpert.news/pt/article/recursive-agent-harnesses-multi-level-agent-decomposition-at-production-scale ai|expert pt-BR 2026-06-12T03:40:10.476Z Agente Recursivo com Aplicação Obtém 89% de Precisão em Tarefas de Código de Longo Contexto https://aiexpert.news/es/article/recursive-agent-harnesses-multi-level-agent-decomposition-at-production-scale ai|expert es 2026-06-12T03:40:10.476Z Agente recursivo alcanza un 89% de precisión en tareas de código de larga duración https://aiexpert.news/en/article/46-of-ai-generated-code-fixes-rejected-new-dataset ai|expert en 2026-06-12T02:44:10.450Z Half of AI-Generated Code Fixes Fail Human Review https://aiexpert.news/pt/article/46-of-ai-generated-code-fixes-rejected-new-dataset ai|expert pt-BR 2026-06-12T02:44:10.450Z Metade das Correções de Código Geradas por IA Não Passam na Revisão Humana https://aiexpert.news/es/article/46-of-ai-generated-code-fixes-rejected-new-dataset ai|expert es 2026-06-12T02:44:10.450Z La mitad de las correcciones de código generadas por IA fallan en la revisión humana https://aiexpert.news/en/article/direct-allocating-test-time-compute-for-embodied-ai-agents ai|expert en 2026-06-12T01:45:19.217Z DIRECT cuts embodied AI latency 65% with dynamic planner routing https://aiexpert.news/pt/article/direct-allocating-test-time-compute-for-embodied-ai-agents ai|expert pt-BR 2026-06-12T01:45:19.217Z DIRECT reduz a latência da IA encarnada em 65% com roteamento dinâmico do planejador https://aiexpert.news/es/article/direct-allocating-test-time-compute-for-embodied-ai-agents ai|expert es 2026-06-12T01:45:19.217Z DIRECT reduce la latencia de IA encarnada en un 65% con enrutamiento dinámico de planificador https://aiexpert.news/en/article/claude-fable-5-autonomously-built-cors-servers-screenshot-tools-and-template-pat ai|expert en 2026-06-12T01:14:10.532Z Claude Fable 5 Autonomously Patched Code and Cost $110 in a Day https://aiexpert.news/pt/article/claude-fable-5-autonomously-built-cors-servers-screenshot-tools-and-template-pat ai|expert pt-BR 2026-06-12T01:14:10.532Z Claude Fable 5 Corrigiu Código de Forma Autônoma e Custou US$ 110 em um Dia https://aiexpert.news/es/article/claude-fable-5-autonomously-built-cors-servers-screenshot-tools-and-template-pat ai|expert es 2026-06-12T01:14:10.532Z Claude Fable 5 parcheó código de forma autónoma y costó 110 dólares en un día https://aiexpert.news/en/article/modsleuth-audits-llms-invisible-upstream-dependencies ai|expert en 2026-06-12T00:00:10.464Z New Tool Finds 1,060 Hidden Training Dependencies Across Major LLMs https://aiexpert.news/pt/article/modsleuth-audits-llms-invisible-upstream-dependencies ai|expert pt-BR 2026-06-12T00:00:10.464Z Nova Ferramenta Identifica 1.060 Dependências de Treinamento Ocultas em Principais LLMs https://aiexpert.news/es/article/modsleuth-audits-llms-invisible-upstream-dependencies ai|expert es 2026-06-12T00:00:10.464Z Nueva Herramienta Detecta 1,060 Dependencias de Entrenamiento Ocultas en Principales LLM https://aiexpert.news/en/article/appo-credit-assignment-at-decision-steps-not-tool-boundaries-for-agentic-rl ai|expert en 2026-06-11T22:02:10.463Z Token-Level Branching Offers Faster LLM Agent Training Without Budget Expansion https://aiexpert.news/pt/article/appo-credit-assignment-at-decision-steps-not-tool-boundaries-for-agentic-rl ai|expert pt-BR 2026-06-11T22:02:10.463Z Ramificação em Nível de Token Oferece Treinamento Mais Rápido para Agente de LLM Sem Expansão de Orçamento https://aiexpert.news/es/article/appo-credit-assignment-at-decision-steps-not-tool-boundaries-for-agentic-rl ai|expert es 2026-06-11T22:02:10.463Z Rama a Nivel de Token Ofrece Entrenamiento Más Rápido de Agente LLM Sin Ampliación del Presupuesto https://aiexpert.news/en/article/why-orbital-data-centers-will-break-silicon-valleys-scaling-playbook ai|expert en 2026-06-11T21:28:10.405Z Orbital Data Centers Face an 80-Square-Meter Cooling Wall https://aiexpert.news/pt/article/why-orbital-data-centers-will-break-silicon-valleys-scaling-playbook ai|expert pt-BR 2026-06-11T21:28:10.405Z Centros de Dados Orbitais Enfrentam um Muro de Resfriamento de 80 Metros Quadrados https://aiexpert.news/es/article/why-orbital-data-centers-will-break-silicon-valleys-scaling-playbook ai|expert es 2026-06-11T21:28:10.405Z Centros de Datos Orbitales Fracasan ante el Muro de Enfriamiento de 80 Metros Cuadrados https://aiexpert.news/en/article/recoverable-routing-cuts-vlm-inference-costno-token-permanence ai|expert en 2026-06-11T20:56:10.358Z Token Recovery Closes Accuracy Gap While Halving VLM Inference Compute https://aiexpert.news/pt/article/recoverable-routing-cuts-vlm-inference-costno-token-permanence ai|expert pt-BR 2026-06-11T20:56:10.358Z Recuperação de Token Fecha a Fossa de Precisão e Diminui a Computação de Inferência de VLM pela Metade https://aiexpert.news/es/article/recoverable-routing-cuts-vlm-inference-costno-token-permanence ai|expert es 2026-06-11T20:56:10.358Z Recuperación de Tokens Cierra Brecha de Precisión Mientras Reduce a la Mitad el Cálculo de Inferencia de VLM https://aiexpert.news/en/article/how-ergo-hestia-cut-model-time-to-market-with-databricks ai|expert en 2026-06-11T20:20:10.419Z Databricks Lakebase Eliminates 10× Latency Spikes at ERGO Hestia https://aiexpert.news/pt/article/how-ergo-hestia-cut-model-time-to-market-with-databricks ai|expert pt-BR 2026-06-11T20:20:10.419Z Databricks Lakebase Elimina Picos de Latência 10× na ERGO Hestia https://aiexpert.news/es/article/how-ergo-hestia-cut-model-time-to-market-with-databricks ai|expert es 2026-06-11T20:20:10.419Z Databricks Lakebase Elimina los Picos de Latencia 10× en ERGO Hestia https://aiexpert.news/en/article/tahoe-error-driven-text-to-sql-for-production-schemas ai|expert en 2026-06-11T19:20:10.259Z Tahoe Text-to-SQL System Cuts Compiler Feedback by 96% https://aiexpert.news/pt/article/tahoe-error-driven-text-to-sql-for-production-schemas ai|expert pt-BR 2026-06-11T19:20:10.259Z Sistema Tahoe Text-to-SQL Reduz Retroalimentação do Compilador em 96% https://aiexpert.news/es/article/tahoe-error-driven-text-to-sql-for-production-schemas ai|expert es 2026-06-11T19:20:10.259Z Sistema Tahoe de Text-to-SQL Reduce el Retroalimentación del Compilador en un 96% https://aiexpert.news/en/article/data-fluency-the-prerequisite-for-ai-scaling ai|expert en 2026-06-11T18:48:10.231Z Aer Lingus Spent 18 Months Rebuilding Data Before AI https://aiexpert.news/pt/article/data-fluency-the-prerequisite-for-ai-scaling ai|expert pt-BR 2026-06-11T18:48:10.231Z Aer Lingus Gastou 18 Meses Reconstruindo Dados Antes de IA https://aiexpert.news/es/article/data-fluency-the-prerequisite-for-ai-scaling ai|expert es 2026-06-11T18:48:10.231Z Aer Lingus Pasó 18 Meses Reconstruyendo Datos Antes de la IA https://aiexpert.news/en/article/production-guidance-on-context-engineering-for-large-scale-ai-systems ai|expert en 2026-06-11T18:16:10.538Z Agent Economics Hinge on Kafka Pipelines, Not Context Windows https://aiexpert.news/pt/article/production-guidance-on-context-engineering-for-large-scale-ai-systems ai|expert pt-BR 2026-06-11T18:16:10.538Z Economia de Agentes Gira em Pipelines do Kafka, Não em Janelas de Contexto https://aiexpert.news/es/article/production-guidance-on-context-engineering-for-large-scale-ai-systems ai|expert es 2026-06-11T18:16:10.538Z La economía de los agentes depende de las tuberías de Kafka, no de las ventanas de contexto https://aiexpert.news/en/article/microsofts-postgresql-extension-enables-durable-execution-for-agent-workflows ai|expert en 2026-06-11T04:46:10.478Z Microsoft Adds Workflow Orchestration to PostgreSQL https://aiexpert.news/pt/article/microsofts-postgresql-extension-enables-durable-execution-for-agent-workflows ai|expert pt-BR 2026-06-11T04:46:10.478Z Microsoft Adiciona Orquestração de Fluxo de Trabalho ao PostgreSQL https://aiexpert.news/es/article/microsofts-postgresql-extension-enables-durable-execution-for-agent-workflows ai|expert es 2026-06-11T04:46:10.478Z Microsoft Incorpora Orquestación de Flujos de Trabajo a PostgreSQL https://aiexpert.news/en/article/diffusiongemma-4x-faster-text-generation-with-optimized-inference-scheduling ai|expert en 2026-06-11T04:12:10.172Z Google's DiffusionGemma Hits 1,000 Tokens Per Second https://aiexpert.news/pt/article/diffusiongemma-4x-faster-text-generation-with-optimized-inference-scheduling ai|expert pt-BR 2026-06-11T04:12:10.172Z Google's DiffusionGemma Atinge 1.000 Tokens por Segundo https://aiexpert.news/es/article/diffusiongemma-4x-faster-text-generation-with-optimized-inference-scheduling ai|expert es 2026-06-11T04:12:10.172Z Google's DiffusionGemma alcanza 1.000 Tokens por Segundo https://aiexpert.news/en/article/reinforcement-learning-optimizes-conversation-dynamics-in-full-duplex-voice-agen ai|expert en 2026-06-11T03:14:10.249Z GRPO Cuts Pause-Handling Errors in Full-Duplex Agents Without Semantic Loss https://aiexpert.news/pt/article/reinforcement-learning-optimizes-conversation-dynamics-in-full-duplex-voice-agen ai|expert pt-BR 2026-06-11T03:14:10.249Z GRPO Reduz Erros de Tratamento de Pausas em Agentes Full-Duplex Sem Perda Semântica https://aiexpert.news/es/article/reinforcement-learning-optimizes-conversation-dynamics-in-full-duplex-voice-agen ai|expert es 2026-06-11T03:14:10.249Z GRPO Reduce las Fallas de Manejo de Pausas en Agentes de Full-Duplex Sin Pérdida Semántica https://aiexpert.news/en/article/databricks-ai-serving-platform-model-agnostic-infrastructure-reduces-per-model-e ai|expert en 2026-06-11T02:42:10.678Z Databricks AI Platform Cuts Infrastructure Costs 90 Percent in Migration Cases https://aiexpert.news/pt/article/databricks-ai-serving-platform-model-agnostic-infrastructure-reduces-per-model-e ai|expert pt-BR 2026-06-11T02:42:10.678Z Plataforma de IA da Databricks Reduz Custos de Infraestrutura em 90% nos Casos de Migração https://aiexpert.news/es/article/databricks-ai-serving-platform-model-agnostic-infrastructure-reduces-per-model-e ai|expert es 2026-06-11T02:42:10.678Z Plataforma de IA de Databricks Reduce el Costo de Infraestructura en un 90 Porciento en Casos de Migración https://aiexpert.news/en/article/google-locks-in-3m-tpus-with-intel-by-2028signals-inference-capacity-surge ai|expert en 2026-06-11T02:10:09.940Z Google Orders 3M TPUs From Intel, Breaking TSMC Packaging Monopoly https://aiexpert.news/pt/article/google-locks-in-3m-tpus-with-intel-by-2028signals-inference-capacity-surge ai|expert pt-BR 2026-06-11T02:10:09.940Z Google Contrata 3M TPUs da Intel, Quebrando Monopólio de Empacotamento da TSMC https://aiexpert.news/es/article/google-locks-in-3m-tpus-with-intel-by-2028signals-inference-capacity-surge ai|expert es 2026-06-11T02:10:09.940Z Google ordena 3M TPUs a Intel, rompiendo el monopolio de empaque de TSMC https://aiexpert.news/en/article/samsung-supermicro-deploy-50mw-floating-ai-data-centers ai|expert en 2026-06-11T01:38:10.078Z 50 MW Floating AI Data Center Wins Design Approval Without Customer Contract https://aiexpert.news/pt/article/samsung-supermicro-deploy-50mw-floating-ai-data-centers ai|expert pt-BR 2026-06-11T01:38:10.078Z Centro de Dados Flutuante de 50 MW de IA Ganha Aprovação de Projeto Sem Contrato de Cliente https://aiexpert.news/es/article/samsung-supermicro-deploy-50mw-floating-ai-data-centers ai|expert es 2026-06-11T01:38:10.078Z Centro de Datos AI Flotante de 50 MW Obtiene Aprobación de Diseño Sin Contrato de Cliente https://aiexpert.news/en/article/azure-api-management-adds-unified-multi-model-serving-gateway ai|expert en 2026-06-11T01:01:04.951Z Azure's Unified Model API masks silent streaming failures https://aiexpert.news/pt/article/azure-api-management-adds-unified-multi-model-serving-gateway ai|expert pt-BR 2026-06-11T01:01:04.951Z API de Modelo Unificado do Azure mascara falhas de streaming silenciosas https://aiexpert.news/es/article/azure-api-management-adds-unified-multi-model-serving-gateway ai|expert es 2026-06-11T01:01:04.951Z La API de Modelo Unificada de Azure enmascara fallos de transmisión silenciosa https://aiexpert.news/en/article/phase-diagram-for-multimodal-learning-align-vs-predict ai|expert en 2026-06-11T00:30:18.525Z Kamai's Phase Diagram Predicts Multimodal Failure Before GPU Commit https://aiexpert.news/pt/article/phase-diagram-for-multimodal-learning-align-vs-predict ai|expert pt-BR 2026-06-11T00:30:18.525Z Diagrama de Fases de Kamai Prevê Falhas Multimodais Antes do Investimento em GPU https://aiexpert.news/es/article/phase-diagram-for-multimodal-learning-align-vs-predict ai|expert es 2026-06-11T00:30:18.525Z Diagrama de Fases de Kamai Predice Fallas Multimodales Antes del Compromiso de GPU https://aiexpert.news/en/article/abc-bench-measuring-llm-biosecurity-risk ai|expert en 2026-06-11T00:00:10.011Z ABC-Bench Shows LLM Agents Now Outperform Expert Biologists on Lab Tasks https://aiexpert.news/pt/article/abc-bench-measuring-llm-biosecurity-risk ai|expert pt-BR 2026-06-11T00:00:10.011Z ABC-Bench Mostra que Agentes de LLM Agora Superam Biólogos Especiais em Tarefas de Laboratório https://aiexpert.news/es/article/abc-bench-measuring-llm-biosecurity-risk ai|expert es 2026-06-11T00:00:10.011Z ABC-Bench Demuestra que los Agentes de LLM Ya Superan a los Biólogos Expertos en Tarefas de Laboratorio https://aiexpert.news/en/article/steering-reasoning-models-by-predicting-future-misbehavior ai|expert en 2026-06-10T15:20:10.039Z FPCG steers reasoning models at test time without retraining https://aiexpert.news/pt/article/steering-reasoning-models-by-predicting-future-misbehavior ai|expert pt-BR 2026-06-10T15:20:10.039Z FPCG dirige modelos de raciocínio em tempo de teste sem retreinamento https://aiexpert.news/es/article/steering-reasoning-models-by-predicting-future-misbehavior ai|expert es 2026-06-10T15:20:10.039Z FPCG dirige modelos de razonamiento en tiempo de prueba sin reentrenamiento https://aiexpert.news/en/article/predictive-signals-improve-steering-of-reasoning-models ai|expert en 2026-06-10T14:40:10.592Z Linear Probes Predict Reasoning-Model Behavior at 64–91% Accuracy https://aiexpert.news/pt/article/predictive-signals-improve-steering-of-reasoning-models ai|expert pt-BR 2026-06-10T14:40:10.592Z Sondas Lineares Atingem 64-91% de Precisão para Modelos de Raciocínio Direcionados https://aiexpert.news/es/article/predictive-signals-improve-steering-of-reasoning-models ai|expert es 2026-06-10T14:40:10.592Z Sondeos Lineales Logran Precisión del 64-91% en Modelos de Razonamiento