Tag
Anthropic is reportedly paying less per GPU than Google for SpaceX compute, with Google paying $920m for 110k GPUs compared to Anthropic's $1.25b for 220k GPUs plus additional capacity, highlighting a significant cost discrepancy.
Nvidia's VP states compute costs now exceed employee costs for his team; Uber confirms by exhausting its 2026 AI coding budget by April due to high token costs.
MiniMax's new m3 model achieves the same score as Opus 4.7 on terminal-bench 2.1 while using 1/20th the compute and cost, attributed to their novel MiniMax Sparse Attention architecture.
A commentary questioning why users cannot run Gemini and Claude Code locally on their own GPUs, implying compute cost constraints are limiting access to these AI models.
A user shares their experience spending ¥3,000 worth of tokens to create an AI video, suggesting that creativity, rather than compute power, is the true bottleneck.
This position paper argues that AI agents are a production technology rather than a labor input, proposing a 'Compute-Anchored Wage' bound where human wages are determined by compute capital costs rather than labor supply elasticity.