OpenAI trails Anthropic as losses deepen and Altman pauses frontier AI training

Widening losses and intensifying competition coincide with a pause in frontier reinforcement-learning training as openai strengthens its safety controls.