TL;DRGoogle has introduced new lightweight versions of its Gemini language model, specifically the 3.6 Flash and 3.5 Flash-Lite iterations. Meanwhile, development of the more powerful Pro variant continues to encounter difficulties.
Global EconomyForeign PolicyArtificial Intelligence
TL;DRAt a world conference in Shanghai, China emphasized its artificial intelligence ambitions while narrowing the technological gap with the United States. Additionally, Beijing is increasingly focusing on open-source solutions to establish partnerships with the Global South.
TL;DRGoogle is developing proprietary chips that feature integrated Gemini model architectures. This hardware aims to accelerate the execution of artificial intelligence applications and improve overall operational efficiency.
Companies & MarketsArtificial IntelligenceIT & Internet
TL;DRMicrosoft is extending its existing collaboration with the French company Mistral AI. The agreement involves a multi-billion dollar investment intended to strengthen artificial intelligence infrastructure in Europe.
TL;DRThe US Air Force and DARPA are currently testing modified F-16 fighter jets capable of autonomous flight in Florida. The research program aims to evaluate autonomous flight technologies within this specific aircraft model.
TL;DRMexican President Claudia Sheinbaum met with Donald Trump at the conclusion of the World Cup. During the course of the tournament, three journalists were murdered in Mexico.
TL;DRTwo million people in Spain are celebrating the national team's World Cup victory, while Andy Burnham has delivered his first address as the British Prime Minister. Additionally, German Chancellor Friedrich Merz has announced a cabinet reshuffle.
TL;DRFIFA has launched an investigation into the Argentine national team following incidents that occurred after the World Cup final. A concurrent online petition calling for the team's exclusion from the tournament has reportedly reached over 23 million signatures.
TL;DRA report from euronews suggests that intentionally introducing restricted content into AI systems may serve as a method to prevent AI-supported cyberattacks. This approach aims to trigger safety limitations to disrupt potentially malicious operations.