The latest report from The Information provides insights into Apple’s in-house AI chatbot development project, codenamed Ajax GPT. According to the report, this project is currently consuming millions of dollars daily for the development and training of a massive model encompassing over 200,000 parameters.
Internally referred to as Foundational Models, this advanced AI model is being developed by a team of approximately 16 individuals. Some members of this team are former engineers from Google, with John Giannandrea, an experienced AI executive, overseeing their operations. Giannandrea, who previously worked at Google, brings valuable expertise to steer this talented team.
The report also reveals that Apple has separate teams dedicated to developing AI for specific purposes. For instance, there is a team focused on AI for image generation, while another team is working on multimodal AI that can analyze and process various data types, including images, videos, and text. The potential applications for these advancements range from enhancing AppleCare’s customer support to integrating them as features within Siri.
TLDR: The Information’s report sheds light on Apple’s internal project Ajax GPT, a multimillion-dollar AI chatbot development endeavor. Led by former Google executive John Giannandrea, the team is building Foundational Models utilizing over 200,000 parameters. Apple also has specialized teams working on AI for image generation and multimodal analysis. The possibilities for implementing these advancements include bolstering customer support and enhancing Siri.