Open-weights, local inference, and first-party hyperscaler models confront frontier labs
Open-weights, local inference, and first-party hyperscaler models are challenging frontier labs like OpenAI and Anthropic, with newly crowned AI models spending only 41 days atop leaderboards before b…