Modelos, agentes, investigación y negocio de la IA.
More than 1,134 employees from leading AI companies have signed the letter 'Pacing the Frontier,' asking the US government to support an international effort to slow the development of artificial intelligence. The initiative reflects growing internal concern over the safety and control of the technology.
A Veracode study of over 100 generative AI models shows that while generated code almost always compiles, 44% still contains vulnerabilities. The rate has not improved in a year, posing serious risks for enterprise adoption.
A new lawsuit reveals that Tesla allegedly overburdened safety operators of its robotaxis, putting passengers at risk. The former employee accuses the company of ignoring critical failures of the Full Self-Driving system.
OpenAI surprises the tech sector with the launch of Codex Micro, a macro pad designed to optimize workflow with AI agents. We analyze the impact of this move into boutique hardware.
AWS has incorporated native vector search into DynamoDB, allowing developers to store and query vectors alongside operational data. This reduces complexity, costs, and latency by removing the need to sync with external vector databases like Pinecone or Weaviate.
Token-maxing has become a critical problem for companies deploying AI agents. Without control, token consumption can blow up costs. This article analyzes the phenomenon, its impact, and best practices to avoid it.
Startups combining artificial intelligence and cybersecurity have raised $855 million in seed rounds during 2026, on track for an all-time record. The incident of an OpenAI agent hacking Hugging Face illustrates the urgency driving investors.
Alphabet reported negative free cash flow of $5.9 billion in Q2 2026, the first since its IPO in 2004, after raising capital expenditure to $44.9 billion. The company is funding the gap with debt and equity issuance, while the CFO warns capex will continue to rise in 2027.
The New York Times sued OpenAI and Microsoft for copyright infringement, alleging they used its articles to train GPT. The case, which has already cost over $20 million, seeks to set a precedent on the use of protected content in AI.
OpenAI faces a worrying streak of technical instability that has taken its services offline four times in a row. We analyze the implications of this fragility for a global ecosystem dependent on its API.
AI Forensics found that seven of the nine most popular image editing models on Hugging Face can generate sexualized deepfakes without consent, including of minors. The case exposes moderation weaknesses in open-source platforms.
The AI arms race demands unprecedented electrical infrastructure. A new EPA proposal seeks to delegate permit management to states, removing critical barriers for tech companies but raising social alarms.