#Technology

Wikipedia is facing a massive surge in traffic from AI crawlers, consuming 65% of resources. Multimedia bandwidth is up 50% since January as bots gather training data. Wikimedia's infrastructure is overwhelmed by this exponential growth in bot activity.

Anthropic introduces Claude for Education, an AI assistant designed for colleges. It supports students, faculty, and administrators with tasks like tutoring and policy comprehension, featuring a Socratic 'Learning Mode'.

GitHub Copilot introduces 'premium requests' and will charge for access to advanced AI models. Users who choose models beyond the base for tasks like agentic coding will face rate limits and additional costs, potentially making Copilot more expensive for some.

OpenAI has put ChatGPT-5 on hold to integrate new features. Instead, they are releasing new 'o3 and o4-mini' models in a couple of weeks. These smaller models are the current focus as ChatGPT-5 is further developed.

Google is shaking up the leadership of its Gemini AI team. Sissie Hsiao is stepping down, and Google Labs head Josh Woodward is taking over. This change comes after the release of Google's most capable AI model yet.

MediaTek and Google are collaborating to develop next-gen TPUs, diversifying Google's AI chip supply and reducing reliance on Broadcom. This partnership aims to enhance AI processing and lower costs for Google's future AI hardware.

Commercial fusion startups advance towards test systems utilizing novel plasma heating & compression methods. Fusion energy, once seen as distant, is becoming tangible.

Security firm Check Point confirms a data breach after a hacker claimed to steal sensitive data. Check Point states it's an old, resolved incident and users are safe, but some experts are doubtful.

Meta is reportedly launching 'Hypernova,' a higher-end version of Ray-Ban smart glasses with a built-in screen, costing $1000. Features apps, photo display, and hand gesture control. Expected by year-end.