Apple vs. Google AI: MLX Targets machine learning, while Gemini aims for multimodality
Delhi : Apple has released MLX, a model library and AI framework, which together should enable the company's CPUs to perform AI functions on MacBooks and applications. Developers may create models to run on the Apple Silicon platform using the framework and deep learning model library that Apple's machine learning research team provided. These are available on GitHub and PyPI.
It is said that MLX is best suited for developers working on machine learning models, however it is not a tool for end users. The architecture of the framework appears to have been influenced by PyTorch, ArrayFire, Jax, and other sources. Additionally, it supports unified memory models, which eliminates the need for data copies when operating arrays in shared memory on connected devices.
Composable function transformations, dynamic computational network construction, and lazy calculations with arrays materialising only when needed are some additional special aspects of Apple's AI framework. The framework is not anticipated to compete with Microsoft's and Google's AI apps, though, as it is primarily focused on machine learning applications. Apple has not kept up with the rest of the pack in the AI race.
Google revealed the release of Gemini, its most recent and potent generative AI model. Gemini is anticipated to go up against OpenAI's GPT-4. The LLM belongs to the type of natively multimodal models.
As a result, it can handle text, video, photos, audio, and code. Google produced a video demonstrating the AI model's capabilities, demonstrating that the AI could effortlessly detect drawings and actual items, as well as concepts linked with such objects. Gemini can also grasp programming languages like as Go, Python, C++, and Java.
Gemini will be offered in three sizes: Ultra, Pro, and Nano. Gemini Ultra will be used largely in data centres for more sophisticated activities. Individuals and small businesses could use the Pro version, while the Nano version was designed to operate on devices like as Google's Pixel smartphone line. Bard Advanced, a new version of Bard powered by Gemini, is also in the works, according to the company.
Following the debut of ChatGPT in 2022, Google has been playing catch-up with its competitors. However, Gemini goes far further in levelling the playing field than Google. Its success is believed to be determined by how successfully Google integrates it with other products in its portfolio, including as its search engine and workplace apps.
