New AI frameworks hint at Apple’s generative AI roadmap

Apple Inc. has quietly launched new machine studying frameworks and mannequin libraries, MLX and MLX Knowledge, signaling a rising curiosity in pursuing generative AI functions, in line with a latest The Verge report. The releases come as a shock given Apple’s traditionally conservative method to AI expertise.

MLX is a machine studying framework optimized to run effectively on Apple Silicon chips. It makes it simple for builders to construct AI fashions that may run on MacBook and different Apple gadgets. MLX Knowledge is a supplementary knowledge loading framework.

In response to Apple’s GitHub page, MLX attracts inspiration from present frameworks like PyTorch and Jax however is uniquely designed for seamless integration throughout Apple {hardware}, using shared reminiscence for quicker mannequin coaching. Builders have famous that MLX seems able to coaching complicated AI fashions corresponding to Meta’s Llama and Steady Diffusion.

Will Apple pursue the ever-expanding generative AI area?

The launch suggests Apple could also be poised to pursue the favored generative AI area, together with textual content, picture, and video era, an space firms like Microsoft and Google have quickly expanded into lately.

Whereas Apple has utilized AI expertise in its merchandise for years, it has historically averted labeling it as such, preferring phrases like “machine studying.” The corporate additionally seems to have accelerated its AI improvement efforts this previous September to guage foundational fashions for broader implementation.

Trade analysts theorize Apple might leverage its new frameworks to introduce modern generative AI options throughout its providers and gadgets — though the traditionally privacy-centric firm will doubtless assess moral implications intently first. Apple declined requests for additional touch upon its particular MLX, MLX Knowledge, and generative AI plans.

Generative AI refers to a category of synthetic intelligence capabilities targeted on truly creating or “producing” new content material reasonably than simply analyzing knowledge or responding to enter. The most typical kinds of generative AI generate artificial textual content, pictures, audio, and video that intently mimic human-produced artistic works.

Featured Picture Credit score: Picture by Armand Valendez; Pexels

Radek Zielinski

Radek Zielinski is an skilled expertise and monetary journalist with a ardour for cybersecurity and futurology.