Microsoft is steering away from creating uncontrollable general-purpose AI and turning toward Humanist Superintelligence, a new approach designed to prioritize human oversight and social benefit.
A 1.5 kilometre stretch of the A10 near Paris is now delivering up to 300 kW of wireless charging to electric vehicles in real traffic conditions. Early results suggest dynamic charging could reduce battery size, vehicle weight, and operating costs, although nationwide deployment remains complex and expensive.
Nvidia CEO Jensen Huang believes China is only a fraction behind the United States in AI development and warns that restrictive chip policies risk pushing half of the world’s developers out of the American tech ecosystem. His comments raise tough questions about how the US plans to stay ahead in a race that is no longer one sided.
New data from Proofpoint shows that Gen AI tools and autonomous agents are creating a fresh category of insider risk. Companies now face growing exposure from unsupervised AI access, careless employees, and sensitive data flowing into public models. A shift toward behavioural analytics and unified security controls is becoming essential.
OpenAI has added a feature that allows users to interrupt ChatGPT mid-response and adjust the prompt without restarting the entire conversation. The change offers a more practical way to guide long, detailed answers and reduces wasted time when the model takes several minutes to respond.
Perplexity has updated its Comet Assistant to handle multiple tabs at once, manage long chains of tasks, and take action in the browser with user permission. The goal is simple. Give people an assistant that can stay with complicated work instead of breaking down the moment things get busy.
Google Cloud has upgraded its Vertex AI Agent Builder with new tools that speed up development and strengthen security. The updates introduce a smarter Agent Development Kit, prebuilt plugins, a self-healing capability, expanded language support, improved observability, and new protection layers to help developers move from prototypes to production more reliably.
Internal projections suggest Meta makes around $16 billion a year from adverts linked to scams and prohibited products. Despite years of public crackdowns, leaked documents indicate Meta allows high-risk ads to run until the probability of fraud hits 95 percent, raising questions about whether the company has any real incentive to shut down bad actors.
