Dario Amodei argues that frontier AI labs should deliberately slow the pace of model capability improvements. The article frames this proposal against Beijing’s speed-first AI safety regime and describes AI safety as an urgent debate.
SemiAnalysis measures usage limits across AI subscription plans and introduces a dashboard tracking plans from Anthropic, OpenAI, Meta, SpaceXAI, Cursor, Cognition, Z.ai, MiniMax, and Moonshot. Its rough estimate says Anthropic subscriptions generate 10% of revenue while consuming more than 40% of inference compute, lowering blended revenue per megawatt by about $36 million. Experiments isolate input, cache-write, cache-read, and output tokens to estimate how plans use limits across models.
MiniMax Group Inc. announced its unaudited financial results for the first half of 2026. The Hong Kong-based company, listed on HKEX as 00100, made the announcement on August 26, 2026.
Shanghai-based MiniMax debuts M3, an AI model designed for long and complex coding tasks. The company says M3 processes data five times faster than its predecessor while reducing inference costs.
