Apple Makes First Move In AI, Even as Google Launches Gemini Model
The Gemini LLM will be available in Ultra, Pro, and Nano suites based on use cases and device applications.
- Apple’s machine learning team has released a new AI and ML framework and model libraries for Apple Silicon.
- Google has launched its latest and most powerful artificial intelligence model, Gemini, which could potentially compete with OpenAI’s GPT-4.
Apple has launched MLX, its AI framework, and MLX Data, a model library, both expected to run the company’s chips for AI operations on MacBooks and apps. The framework and deep learning model library, released by Apple’s machine learning research team, can be used by developers to develop models to operate on the Apple Silicon platform and is accessible through GitHub and PyPI.
While MLX is not a tool for end consumers, it is stated to be ideal for developers working on machine learning models. The framework design is apparently inspired by Jax, ArrayFire, PyTorch, and others. It also supports unified memory models, allowing arrays in shared memory to run on associated devices without having to use copies of the data.
Some other unique features of Apple’s AI framework include lazy computations with arrays only materializing when required, dynamic building of computational graphs, and composable function transformations. However, the framework is largely focused on ML applications and is not expected to compete with the AI applications of Microsoft and Google. Apple has remained a laggard in the AI race so far.
See More: Elon Musk’s xAI is Raising $1B as OpenAI’s Advanced Models Integrate With Copilot
Google announced the launch of Gemini, its latest and most powerful generative AI model. Gemini is expected to compete with OpenAI’s GPT-4. The LLM is of the natively multimodal model type.
Consequently, it can work with text, video, images, audio, and code. Google released a video that demonstrated the capabilities of the AI model, which showed that the AI could easily identify drawings and physical objects, in addition to concepts associated with such objects. Gemini can also understand coding languages such as Go, Python, C++, and Java, among others.
Gemini will be available to users in three variants — Ultra, Pro, and Nano. Data centers will primarily use Gemini Ultra for tasks of higher complexity. Pro, the mid-range version could be used by individuals and smaller organizations, while the Nano version has been created to run on devices like Google’s Pixel smartphone range. The company has also stated plans to release a new version of Bard, powered by Gemini, called Bard Advanced.
Google has been playing catchup with its competitors following the launch of ChatGPT in 2022. However, Gemini takes Google much further in leveling the playing field. Its success is expected to depend on how well Google manages to integrate it with other products in its portfolio, including its search engine, workspace apps, and others.
What do you think about Big Tech investments in artificial intelligence? Let us know your thoughts on LinkedIn, X, or Facebook. We’d love to hear from you!
Image source: Shutterstock
LATEST NEWS STORIES
