Apple Teases New Generative AI Features That Could Soon Become Part Of The iPhone

For a while now, tech giant Apple has been fairly hush about its future plans regarding the world of generative AI. Now, we’re hearing more on this front as the leading iPhone maker teased a host of AI features that might soon become a part of its devices.

The company was certainly ambitious on this front in regards to making AI function locally across its long list of devices. Moreover, we heard about so many researchers from the tech giant testing an OpenELM that featured an array of four tiny language models across its Hugging Face library.

The company added how the OpenELM is doing a great job in terms of productivity for several text-based tasks including email generation. Such models are not only open source but they’re set up for use by developers as well.

It entails four sizes which include 270, 450, 1.1 billion, and also 3 billion parameters. This latter term is reserved for the figure of variables that every model understands when deciding on a host of datasets for training.

For instance, we’ve got the software giant rolling out the Phi-3 model while Google’s Gemma is giving rise to another 2 billion variant.

Smaller scale models are not only cheaper when they run but also offer greater optimization when functioning across different devices such as the usual laptops and mobile phones.

But Apple’s CEO seems confident about Generative AI coming to Apple as he teased more about the launch starting in February and how Apple might be taking part in excessive spending of its resources to achieve certain specifics on what actual AI might appear such as in the iPhone collection.

It wouldn’t be wrong to mention that we’ve seen the firm roll out AI models in the past but we’re yet to see the launch of AI foundation models linked to commercial use which competitors entail.

This past December, Apple rolled out its MLX which is a machine learning endeavor that makes it so much simpler for these kinds of variants to run better through Apple Silicon. It even sheds light on the MGIF which happens to be a new image editor that enables users to amend pictures through prompts.

In the same manner, we’ve heard more about Ferret-UI that might be utilized for things like smartphone navigation. Apple is said to working on similar endeavors like code completion that’s quite like GitHub’s Copilot.

But despite all such releases from the tech giant, the firm continues to reach out to rivals to have their models on Apple’s products.

Image: DIW-Aigen

Read next: US Congress Approves Full TikTok Ban Across The Country
Previous Post Next Post