> M4 has Apple’s fastest Neural Engine ever, capable of up to 38 trillion operations per second, which is faster than the neural processing unit of any AI PC today.
How useful this is for free libraries? Can you invoke it from your Python or C++ code in a straightforward manner? Does it not rely on proprietary drivers only available for their own OS?
It's a piece of hardware so it'll have their own driver but you can access it through the OS APIs just like any other piece of hardware. What's new is the improvement in speed, Apple has exposed their NPUs for years. You can e.g. run Stable Diffusion on them.
How useful this is for free libraries? Can you invoke it from your Python or C++ code in a straightforward manner? Does it not rely on proprietary drivers only available for their own OS?