https://aiexpert.news/en/article/sovereignpa-bench-measuring-personal-agent-trustworthiness ai|expert en 2026-07-07T12:52:42.405Z SovereignPA-Bench Measures Whether AI Agents Protect User Boundaries https://aiexpert.news/pt/article/sovereignpa-bench-measuring-personal-agent-trustworthiness ai|expert pt-BR 2026-07-07T12:52:42.405Z SovereignPA-Bench Mede se Agentes de IA Protegem Limites do Usuário https://aiexpert.news/es/article/sovereignpa-bench-measuring-personal-agent-trustworthiness ai|expert es 2026-07-07T12:52:42.405Z SovereignPA-Bench Mide si los Agentes de IA Protegen los Límites del Usuario https://aiexpert.news/en/article/cortex-bridging-planning-and-execution-in-long-horizon-embodied-tasks ai|expert en 2026-07-07T12:20:42.415Z Dual Planning Loop Solves Semantic Gap in Hierarchical Robots https://aiexpert.news/pt/article/cortex-bridging-planning-and-execution-in-long-horizon-embodied-tasks ai|expert pt-BR 2026-07-07T12:20:42.415Z Loop Duplo de Planejamento Resolve Fossa Semântica em Robôs Hierárquicos https://aiexpert.news/es/article/cortex-bridging-planning-and-execution-in-long-horizon-embodied-tasks ai|expert es 2026-07-07T12:20:42.415Z Bucle de Planificación Dual Resuelve el Vacío Semántico en Robots Jerárquicos https://aiexpert.news/en/article/gap-framework-bridging-interpretability-and-adaptation-in-commercial-robot-autom ai|expert en 2026-07-07T10:34:42.239Z Graph-as-Policy Framework Achieves 0.93-0.99 Success Rate on Robot Manipulation https://aiexpert.news/pt/article/gap-framework-bridging-interpretability-and-adaptation-in-commercial-robot-autom ai|expert pt-BR 2026-07-07T10:34:42.239Z Quadro Graph-as-Policy Alcança Taxas de Sucesso de 0,93-0,99 em Manipulação de Robôs https://aiexpert.news/es/article/gap-framework-bridging-interpretability-and-adaptation-in-commercial-robot-autom ai|expert es 2026-07-07T10:34:42.239Z Marco Graph-as-Policy Logra Tasas de Éxito del 0.93 al 0.99 en Manipulación de Robots https://aiexpert.news/en/article/searchgen-20k-bridging-visual-generators-knowledge-gap-with-search-integration ai|expert en 2026-07-07T10:20:42.230Z SearchGen-20K Teaches Visual Generators When to Search https://aiexpert.news/pt/article/searchgen-20k-bridging-visual-generators-knowledge-gap-with-search-integration ai|expert pt-BR 2026-07-07T10:20:42.230Z SearchGen-20K Ensina Geradores Visuais Quando Pesquisar https://aiexpert.news/es/article/searchgen-20k-bridging-visual-generators-knowledge-gap-with-search-integration ai|expert es 2026-07-07T10:20:42.230Z SearchGen-20K Ensena a Generadores Visuales Cuándo Buscar https://aiexpert.news/en/article/photoroom-shares-production-data-strategy-for-vision-models ai|expert en 2026-07-07T08:30:42.211Z Photoroom Trades Throughput for Encoder Flexibility in Diffusion Training https://aiexpert.news/pt/article/photoroom-shares-production-data-strategy-for-vision-models ai|expert pt-BR 2026-07-07T08:30:42.211Z Photoroom Troca Desempenho pelo Flexibilidade do Codificador na Treinamento de Difusão https://aiexpert.news/es/article/photoroom-shares-production-data-strategy-for-vision-models ai|expert es 2026-07-07T08:30:42.211Z Photoroom Opta por Flexibilidad del Codificador en Entrenamiento de Difusión a Costa de Rendimiento https://aiexpert.news/en/article/compactionrl-extends-long-horizon-agent-learning-beyond-context-windows ai|expert en 2026-07-07T07:58:42.453Z CompactionRL boosts GLM coding agents 5–7 points on benchmarks https://aiexpert.news/pt/article/compactionrl-extends-long-horizon-agent-learning-beyond-context-windows ai|expert pt-BR 2026-07-07T07:58:42.453Z CompactionRL impulsiona agentes de codificação GLM em 5–7 pontos nos benchmarks https://aiexpert.news/es/article/compactionrl-extends-long-horizon-agent-learning-beyond-context-windows ai|expert es 2026-07-07T07:58:42.453Z CompactionRL impulsa a los agentes de codificación GLM 5–7 puntos en benchmarks https://aiexpert.news/en/article/llm-as-a-verifier-eval-without-fine-tuning ai|expert en 2026-07-07T07:26:42.391Z New Verification Method Hits 86.5% on Terminal-Bench Without Fine-Tuning https://aiexpert.news/pt/article/llm-as-a-verifier-eval-without-fine-tuning ai|expert pt-BR 2026-07-07T07:26:42.391Z Novo Método de Verificação Atinge 86,5% em Terminal-Bench Sem Ajuste Fino https://aiexpert.news/es/article/llm-as-a-verifier-eval-without-fine-tuning ai|expert es 2026-07-07T07:26:42.391Z Nuevo Método de Verificación Logra 86.5% en Terminal-Bench Sin Ajuste Fino https://aiexpert.news/en/article/cloudflare-workers-cache-edge-cached-inference-cuts-cost ai|expert en 2026-07-07T06:54:42.508Z Cloudflare eliminates GPU charges for cached AI inference with new Workers Cache https://aiexpert.news/pt/article/cloudflare-workers-cache-edge-cached-inference-cuts-cost ai|expert pt-BR 2026-07-07T06:54:42.508Z Cloudflare elimina cobranças de GPU para inferência de IA em cache com novo Workers Cache https://aiexpert.news/es/article/cloudflare-workers-cache-edge-cached-inference-cuts-cost ai|expert es 2026-07-07T06:54:42.508Z Cloudflare elimina cargos de GPU para inferencia de IA en caché con nuevo Workers Cache https://aiexpert.news/en/article/chinese-ai-models-gain-share-as-us-inference-costs-surge ai|expert en 2026-07-07T06:22:42.157Z Chinese AI Models Grab 46 Percent of Token Market Share https://aiexpert.news/pt/article/chinese-ai-models-gain-share-as-us-inference-costs-surge ai|expert pt-BR 2026-07-07T06:22:42.157Z Modelos de IA Chinesa Capturam 46% do Mercado de Tokens https://aiexpert.news/es/article/chinese-ai-models-gain-share-as-us-inference-costs-surge ai|expert es 2026-07-07T06:22:42.157Z Modelos de IA chinos capturan el 46 por ciento del mercado de tokens https://aiexpert.news/en/article/hives-silicon-brain-cuts-inference-costs-80but-details-remain-elusive ai|expert en 2026-07-07T05:50:41.984Z Hive Autonomy's silicon brain cuts construction machine costs 80 percent https://aiexpert.news/pt/article/hives-silicon-brain-cuts-inference-costs-80but-details-remain-elusive ai|expert pt-BR 2026-07-07T05:50:41.984Z Cérebro de silício da Hive Autonomy reduz custos de máquinas de construção em 80% https://aiexpert.news/es/article/hives-silicon-brain-cuts-inference-costs-80but-details-remain-elusive ai|expert es 2026-07-07T05:50:41.984Z El cerebro de silicio de Hive Autonomy reduce el costo de las máquinas de construcción en un 80 por ciento