Modelos, agentes, investigación y negocio de la IA.
The launch of Moonshot AI's Kimi K3 model has shaken global markets, with widespread declines in tech and semiconductor stocks. Investors fear China can compete in AI at lower costs, questioning the sustainability of massive spending by US companies.
Major AI models like ChatGPT and Gemini refuse to criticize repressive governments when asked directly, according to an Oversight Board study. This reveals a geopolitical bias that could limit free expression and access to critical information in regions where it is most needed.
The FBI arrested Zyaire Wilkins, accused of distributing malicious games on Steam to steal cryptocurrency. The case exposes a new attack vector combining social engineering, gaming platforms, and digital wallets.
X has used its Grok model to identify and expel over 4,000 creators from its revenue sharing program for content theft. The new detection capability, three times faster than before, marks a drastic shift in platform moderation and redefines incentives for original creators.
Researchers from three European universities found that access to AI advice collapses the willingness to say 'I don't know' from 44% to 3%, while accuracy drops to a third. Confidence, on the other hand, doubles. The finding has profound implications for decision-making in professional and everyday settings.
China faces a new social challenge: young people prefer relationships with artificial intelligences over forming real families. The government has reacted by banning chatbots from simulating emotional bonds, in an attempt to reverse the declining birth rate.
A class action lawsuit filed in Ontario, Canada, alleges that the SAFER artificial intelligence system, used since 2021 to classify inmate security levels, racially discriminates against Black inmates, disproportionately assigning them to maximum security units. The complaint is supported by data showing that Black people, who make up 5% of the provincial population, represent 27% of inmates in maximum security.
Oak, an Israeli startup, has raised $60 million in seed funding to develop an identity operating system that unifies users, machines, and AI agents. The company aims to solve the chaos of enterprise access, where it's currently impossible to know who or what has permissions at any given moment.
Thinking Machines Lab launched Inkling, an open-weight model with 975 billion parameters that doesn't claim to be the most powerful, but the most honest and efficient. With uncertainty flagging and a reasoning effort dial, it targets companies seeking control and savings.
A study by French and Italian universities shows that using AI makes people less likely to say 'I don't know', increasing confidence even as accuracy drops drastically. Researchers warn of risks to critical thinking, especially in children.
Investor Kevin O'Leary has reignited the debate over data center water consumption by comparing them to golf courses. While the claim is technically correct in aggregate figures, detailed analysis reveals key differences in the type of water used and its local impact.
Enterprise AI is moving from predicting to acting autonomously, creating an accountability gap. To close it, organizations must adopt a 'glass box' approach that allows auditing every decision, inspired by software observability.