How Microsoft Is Building for a World of Metered Intelligence Microsoft is adapting to the rising costs of AI by shifting towards a more pragmatic approach, acknowledging that “unmetered intelligence” will likely be constrained by cost. The company is introducing new hardware and software solutions, like the RTX Spark laptop and enhanced Copilot features, designed to allow developers to utilize AI models more affordably, even if it means using smaller, locally-run models.

• The era of heavily subsidized AI subscriptions, similar to early Uber pricing, is ending as AI labs face profitability pressures.
• Microsoft’s Build conference addressed rising AI costs, with CEO Satya Nadella promising “unmetered intelligence” while acknowledging the need for pragmatic solutions.
• The new RTX Spark laptop allows developers to run medium-sized AI models locally, reducing reliance on token-based cloud services.
• Microsoft’s strategy involves offering flexibility, allowing users to switch between different AI models and providers, including local open-source options.
• Changes in billing, such as GitHub Copilot’s move to token-based pricing, highlight the industry-wide shift towards more direct cost-based models for AI services. Continue reading https://every.to/also-true-for-humans/how-microsoft-is-building-for-a-world-of-metered-intelligence