Modelos, agentes, investigación y negocio de la IA.
Christine Lagarde warned that artificial intelligence could trigger global financial crises and called for international governance similar to nuclear non-proliferation agreements. The statement, made in Venice, is the strongest from a central banker on the systemic risks of AI.
Researchers from NVIDIA GEAR Lab, Carnegie Mellon, and UC Berkeley have developed ENPIRE, a framework that enables AI agents to code and execute training programs for robots without human intervention. The agents taught robotic arms to cut zip ties and insert GPUs into motherboards, operating autonomously overnight.
Anthropic launches a major update to Claude Design that fixes voracious token consumption, adds design system import from GitHub and design files, and positions itself as a brand compliance tool for enterprises. The redesign aims to transform the tool from a flashy prototype into a strategic business asset.
China announced the creation of a global free AI organization for developing countries, just as the G7 discusses restricting access to US models only to trusted partners. The move aims to position Beijing as an inclusive leader against the West's restrictive approach.
Texas attorney Mark Lanier secured a $6 million verdict against Meta and YouTube in the first civil trial for social media addiction. His secret weapon: intensive use of artificial intelligence to analyze documents, prepare interrogations, and synthesize case law in real time.
Elon Musk's legal battle against OpenAI has ended. A federal judge permanently dismissed xAI's lawsuit accusing OpenAI of stealing trade secrets related to the Grok chatbot. The case, which sought to curb competition in the AI market, was rejected for lack of evidence.
Google has introduced its first smart speaker in six years, the new Google Home Speaker, with Gemini integrated and immersive sound. The device aims to regain ground against Amazon and Apple, betting on a more conversational and capable assistant.
Paradromics has implanted its Connexus brain chip in the first participant of an FDA-approved clinical study, marking a milestone against Neuralink. The device aims to restore communication in people with severe paralysis, ushering in a new era for brain-computer interfaces.
AI model chaining uses the output of one model as input for the next, creating a chain of specialization. This technique, increasingly used by experts, refines results and avoids generic responses. We analyze its operation, benefits, and practical applications.
Microsoft is studying incorporating the Chinese model DeepSeek V4 into its Copilot Cowork assistant to lower inference costs. The decision reflects pressure to monetize generative AI and could redefine the company's model strategy.
Two Asian economies have built their growth on AI semiconductors, reaching historic levels of concentration. Analysts warn that any slowdown in demand could trigger a regional crisis with global effects.
Google Cloud announces agreements with HSBC and the UK government to democratize artificial intelligence in enterprises. The strategy focuses on autonomous agents that automate complex processes, marking a milestone in corporate AI adoption.