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

French startup ZML has released an open-source software library called ZML/LLMD designed to optimize the performance of large language models across a variety of hardware architectures. By streamlining how AI models process data on diverse chips, the tool aims to reduce the computational costs associated with running inference. This release provides developers with an alternative to proprietary environments, potentially lowering the financial barriers for organizations deploying AI applications.