Chris Latner, the founder and chief architect of Low Level Virtual Machine (LLVM) and the creator of the SWIFT programming language, along with Tim Davis, the former head of Google’s AI projects, have unveiled a new programming language called Mojo. The language is designed to combine the ease of use for research developments and the rapid creation of prototypes with the ability to create high-performance final products. It achieves this through the use of Python syntax for ease of use, and the ability to compile to machine code, mechanisms for safe memory work, and the use of hardware acceleration of calculations.
While the focus of the project is on machine learning development, it is presented as a general-purpose language, expanding Python capabilities through systematic programming means and suitable for a wide range of tasks, such as high-performance computing, processing, and data transformation. An interesting feature of Mojo is the ability to specify the Emoji symbol “🔥” as an extension for files with code, in addition to the text expansion “.mojo”.
The language is currently in the intensive development stage and is only available for testing through an online interface. Separate assemblies for launching on local systems will be published later, after receiving reviews about the operation of the interactive web-detachment. The initial texts of the compiler, JIT, and other developments related to the project plan to be open after the internal architecture design.
As the Mojo syntax is based on Python and its type is close to C/C++, the project plans to develop tools that simplify the translation of existing projects written in C/C++ and Python to Mojo. They also plan to develop hybrid projects that combine code written in Python and Mojo.
Mojo is calculated to bring a new level of flexibility and performance to programming languages. Its creators believe it has the potential to revolutionize the field of machine learning and beyond.