Apple and Google Gemini: A New Era for iPhone AI
Tech: Google provides its LLM to third-party developers in addition to utilizing Gemini to enhance its apps and services. There are reported discussions between Apple and Google regarding the potential licensing of Gemini for use in the iPhone.
Bloomberg has reported that Apple is currently in negotiations to potentially license Gemini, Google’s generative AI models, for use in upcoming features for the iPhone software in the coming year. Additionally, Apple has held discussions with OpenAI, the company behind Microsoft’s AI capabilities.
Apple is seeking collaborations in the realm of cloud-based generative AI, focusing on applications such as text and image generation for its Gemini project. In addition, Apple is concurrently developing its own on-device AI models and capabilities for integration into the forthcoming iOS 18 release.
The ongoing discussions have yet to determine how the AI agreement will be identified, marking a significant expansion of the current relationship, which primarily revolves around the default search engine between the two companies.
In comparison to other companies in the industry, Google revealed in February that it had entered into a partnership with Samsung to incorporate Gemini power summarization functionality in the Galaxy S24’s notes and voice recording applications, as well as the keyboard. Additionally, Samsung is utilizing Imagen 2 text-to-image diffusion technology for a generative editing feature in the photo gallery app. These features necessitate server-side processing; however, Samsung is also incorporating an on-device iteration of Gemini technology.
Google provides Gemini in three different sizes, where the Pro version is commonly utilized by the majority of both first and third-party applications. The free version of gemini.google.com is supported by Gemini 1.0 Pro, while the paid Gemini Advanced tier utilizes 1.0 Ultra.
Gemini 1.0 is now stable, however, Google introduced a preview of Gemini 1.5 in mid-February with an expanded context window to enhance information absorption. This update is expected to improve the consistency, relevance, and utility of output.
- Gemini Ultra: Largest and most capable model for highly complex tasks
- Gemini Pro: Best model for scaling across a wide range of tasks
- Gemini Nano: Most efficient model for on-device tasks
Article : Link
