GitHub Copilot's Token-Based Billing Sparks Developer Backlash
Microsoft's GitHub Copilot has announced a move to a new token-based billing model, igniting heated debate in the developer community. With Copilot shifting from its previous subscription model to usage-based billing, concerns are spreading that costs will become much harder to predict. TechCrunch reports that "Copilot's golden age may be over," underscoring the end of an era when the tool was available for free or at a flat rate.
The change is believed to be driven by rising operational costs amid surging use of AI coding assistants, but many developers have reacted harshly, with responses along the lines of "this has to be a joke."
Microsoft and Nvidia Reportedly Partnering on an Agentic AI PC
Microsoft and Nvidia are reportedly co-developing a new AI PC that departs from the traditional Copilot approach and runs AI agents locally. According to The Decoder, Nvidia plans to enter the PC market with its own chip as the main processor, and the first Dell and Microsoft Surface products are expected to be announced at Computex and Build next week.
Particularly noteworthy is the fact that the new software will very likely be based on the OpenClaw framework. It is being positioned as the "second arrow" after the Copilot+ PC initiative underperformed, and it is drawing attention as a new approach in which agents handle tasks locally rather than depending on the cloud.
Meta Developing an AI-Powered Pendant
Meta is reportedly working on a wearable pendant with built-in AI capabilities. According to TechCrunch, Meta appears to be making a big bet on AI-powered hardware. For a company already rolling out AI features in its Ray-Ban Meta smart glasses, taking on a new form factor like a pendant could further accelerate the penetration of AI assistants into everyday life. Many details remain unclear, so we will have to wait for future announcements.
Google Launches Always-On AI Assistant "Gemini Spark"
Google has released "Gemini Spark," an always-on AI assistant that automates everyday tasks. According to TechCrunch's hands-on review, it delivers practical results across a wide range of tasks, from summarizing inboxes to planning local events. However, it remains unclear why Google is shipping this as a standalone product, and questions have been raised about where it fits within the company's AI strategy.
Mistral CEO: "Europe Must Build Its Own AI Infrastructure Within Two Years"
An executive at French AI startup Mistral has warned that Europe must build its own AI infrastructure within two years. According to Business Insider, the warning drives home a strong sense of crisis over Europe's lag in the AI field. The executive stressed the importance of having homegrown computing resources and foundational technologies rather than depending on the United States or China — a message that could influence the direction of Europe's AI policy.
Anthropic to Introduce an "AI Fluency Scorecard" for Claude
It has come to light that Anthropic plans to introduce an "AI Fluency Scorecard" for Claude that evaluates users' ability to put AI to work. According to TestingCatalog, it is a mechanism for visualizing an individual's AI proficiency, apparently designed to help users learn and track effective ways of using AI tools. Detailed specifications and release timing are still unknown, but the initiative is drawing attention as a novel approach to quantifying AI literacy.
---