Modelos, agentes, investigación y negocio de la IA.
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.
Databricks has introduced Genie Ontology, a semantic context layer that organizes enterprise knowledge into a living graph. The promise: AI agents that understand the business and provide coherent answers. However, analysts warn that answer verification remains a challenge.
During the G7 summit in Evian-les-Bains, representatives from several member countries proposed to the US administration a 'trusted partners' scheme to facilitate access to its AI models. The initiative seeks to balance national security with technological collaboration among allies.
A US interagency committee approved adding DeepSeek and other Chinese companies to the sanctioned entities list, but the measure is on hold. The delay reflects tensions between national security and commercial interests, and could redefine AI competition.
Belgian research center IMEC has achieved a key breakthrough in 6G chips that solves a critical efficiency problem. This milestone is exactly what Jensen Huang, CEO of Nvidia, needs to materialize his vision of an AI-driven revolution in the telecommunications industry.
Mistral CEO Arthur Mensch has defended the French startup's open source model as a state-control-free option, contrasting with rivals like Anthropic. The statement comes days after Anthropic banned its models in China, intensifying the debate on technological sovereignty and AI access.