Noticias de IA: Two years of OpenAI Academy
Brief editorial de IA: Two years of OpenAI Academy. Noticias y contexto de OpenAI, Google DeepMind, Hugging Face.
OpenAI
Two years of OpenAI Academy
Marking two years of OpenAI Academy and bringing AI skills to even more communities.
Leer en la fuente original : Two years of OpenAI AcademySam Altman’s remarks at the United Nations Security Council
OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.
Leer en la fuente original : Sam Altman’s remarks at the United Nations Security CouncilHarvey turns legal context into stronger drafts with GPT-6 Astra
GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.
Leer en la fuente original : Harvey turns legal context into stronger drafts with GPT-6 AstraHow invideo improves color grading 3x with GPT‑6 Astra
With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.
Leer en la fuente original : How invideo improves color grading 3x with GPT‑6 AstraRingg’s AI agents resolve up to 65% of customer calls with OpenAI
Using GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.
Leer en la fuente original : Ringg’s AI agents resolve up to 65% of customer calls with OpenAIGoogle DeepMind
Advancing Private AI Compute with secure, server-side memory
Introducing private, server-side memory to Private AI Compute for personal AI.
Leer en la fuente original : Advancing Private AI Compute with secure, server-side memoryGemini 3.8 text-to-speech says hello
Leer en la fuente original : Gemini 3.8 text-to-speech says helloHugging Face
How to Use NVIDIA Warp and MjWarp to Accelerate Robotics Simulation and Learning Workflows
Leer en la fuente original : How to Use NVIDIA Warp and MjWarp to Accelerate Robotics Simulation and Learning WorkflowsVoice Cloning with Consent
Leer en la fuente original : Voice Cloning with ConsentStreaming datasets: 100x More Efficient
Leer en la fuente original : Streaming datasets: 100x More Efficienthuggingface_hub v1.0: Five Years of Building the Foundation of Open Machine Learning
Leer en la fuente original : huggingface_hub v1.0: Five Years of Building the Foundation of Open Machine LearningLeRobot v0.4.0: Supercharging OSS Robot Learning
Leer en la fuente original : LeRobot v0.4.0: Supercharging OSS Robot LearningarXiv
Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverse
El estudio analiza los fallos silenciosos en la interacción entre agentes de IA y herramientas en flujos de trabajo biológicos, donde las invocaciones parecen exitosas pero procesan información de forma errónea o incompleta. Esta investigación es relevante al evidenciar deficiencias críticas en la fiabilidad de los sistemas de agentes autónomos que integran diversas herramientas especializadas.
Leer en la fuente original : Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverseHarness as a Language: A Minimalist Agent Framework With Maximal Expressivity
arXiv:2609.26891v1 Announce Type: new Abstract: Modern language-model agents are built around the \textit{agent loop}, where the LLM is placed in an environment exposing a set of tools, and the LLM has full control over the workflow by alternating between tool calls and observin
Leer en la fuente original : Harness as a Language: A Minimalist Agent Framework With Maximal ExpressivityTwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool Agents
TwinCheck es un nuevo método de verificación en tiempo de inferencia que evalúa si una llamada a una herramienta debe ser reemplazada solo cuando existe evidencia que respalde una hipótesis de fallo. Su relevancia radica en prevenir que agentes autónomos se desvíen por acciones erróneas, evitando a su vez que la intervención correctiva introduzca nuevos errores en el proceso.
Leer en la fuente original : TwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool AgentsBuilding Socio-Affective Artificial Intelligence for Interactive Multi-Agent Simulations
arXiv:2609.26927v1 Announce Type: new Abstract: The objective of this article is to provide design principles and a software architecture for enabling interaction between humans and multiple agents in simulated dynamic worlds. This connects the current era of general artificial
Leer en la fuente original : Building Socio-Affective Artificial Intelligence for Interactive Multi-Agent SimulationsWhich Objectives Need a Dial? Predicting Objective Conflict and Covering Trade-offs in Steerable Pluralistic Alignment
arXiv:2609.26929v1 Announce Type: new Abstract: People hold diverse, sometimes conflicting values, so no single aligned model can satisfy everyone. Pluralistic alignment therefore calls for steerable models that can balance competing objectives differently. Multi-Objective Direc
Leer en la fuente original : Which Objectives Need a Dial? Predicting Objective Conflict and Covering Trade-offs in Steerable Pluralistic Alignment