Hot French startup ZML releases free product to speed inference across lots of AI chips

ZML, a French AI startup, has released free software that allows large language models to run across multiple types of AI chips, including those from Nvidia, AMD, Google, Apple, and Intel. The software aims to reduce vendor lock-in and allow companies to use different chips for better cost efficiency and performance. This matters because as AI inference, or the processing of user prompts, becomes increasingly important for AI deployment, optimizing it across diverse hardware could help reduce costs and give organizations more flexibility in their AI infrastructure choices.