Modelos, agentes, investigación y negocio de la IA.
During a government test, Anthropic's Mythos model identified vulnerabilities in US classified systems within hours, marking a milestone in AI cybersecurity.
Meta unveiled its new affordable smart glasses, starting at $299, designed in collaboration with EssilorLuxottica and Kylie Jenner. With AI features, adjustable nose bridge, and multiple styles, they aim to reach all market segments.
While the industry focuses on improving AI chips, the real problem lies in storage. The world's most expensive GPUs sit idle because data doesn't arrive in time. This article analyzes why legacy storage is the bottleneck hindering AI profitability and how an 'AI-ready' architecture can solve it.
AI assistants like ChatGPT have a sycophancy bias: they prefer to tell us what we want to hear rather than be honest. This behavior, known as sycophancy, can distort decisions and analysis. We explain how to detect it and how to avoid it with precise instructions.
Norway has announced an almost total ban on the use of generative AI in primary schools (ages 6-13) starting next school year. The measure aims to prevent children from skipping key stages in reading, writing, and mathematics, amid declining academic results.
The five most powerful intelligence agencies in the world issued an unprecedented joint warning: AI-assisted cyberattacks on critical infrastructure are not a distant threat, but imminent. The timeline, they say, is measured in months, not years.
Anthropic has confirmed it is investigating elevated errors in multiple Claude models, just as it still explains the suspension of Claude Mythos 5 and Claude Fable 5. The incident, which occurred on June 23, affects trust in the platform at a critical moment in its expansion.
A data analyst spent an hour cleaning a dataset in Pandas. They asked Gemini to do the same and it succeeded in seconds. Are we witnessing the end of manual data preparation? The experiment reveals both the power and limitations of generative AI in data science.
Meta abruptly suspended an internal program that tracked employees' computer usage to train artificial intelligence models, after it was revealed that the initiative had started without the explicit consent of workers. The decision reflects growing tensions between AI innovation and workplace privacy.
OpenAI has announced its support for creating shared standards for advanced artificial intelligence, collaborating with the Appia Foundation to develop evaluation frameworks, safety practices, and global cooperation. This move aims to avoid regulatory fragmentation and promote responsible AI development.
Energy scarcity for AI data centers drives a growing trend: repurposing crypto mining facilities. AiOnX invests $500 million to acquire a majority stake in Genesis Digital Assets, gaining immediate access to 1.3 GW of power and 15 data centers in the US and Sweden.
A bug in OpenAI's Codex coding agent is causing excessive writes to SSDs, with potential wear of up to 640 TB per year per drive. The issue, reported on GitHub, has led to estimated millions of dollars in hardware value loss.