i need someone at @OpenAI and @AnthropicAI to teach the models that while prototyping, backwards compatibility is just a bad idea

Jeremy Howard argues that AI model developers at OpenAI and Anthropic should prioritize breaking backwards compatibility during the prototyping phase rather than maintaining it, suggesting that backwards compatibility constraints hinder rapid iteration and experimentation in early-stage AI development. This post advocates for a development philosophy that prioritizes innovation speed and flexibility over maintaining consistency with previous model versions during research and prototyping stages.

Source: Jeremy Howard (X) | 2026-05-22