This week, Apple Intelligence entered beta testing, and the firm has now released a detailed description of the development process for a number of its AI features. An important detail? Two of Apple’s foundation models were developed with CPUs manufactured by Google.
The first comprehensive examination of Apple Intelligence’s growth
Apple often avoids disclosing a lot of information about its internal processes for developing new products. But with AI and ML capabilities, the business has long since made its findings publicly available.
One of the first papers after Apple Intelligence was introduced at WWDC, the most recent publication is named “Apple Intelligence Foundation Language Models.”
It’s not the easiest document to understand because it was written by and for researchers. The chips used to train two of the Apple Intelligence language models are a noteworthy detail, though.
It’s not the easiest document to understand because it was written by and for researchers. The chips used to train two of the Apple Intelligence language models are a noteworthy detail, though.
How are we to interpret this news? Without additional information from Apple, it’s difficult to know what to believe.
Given Apple’s inclination towards internalizing as much labor as possible, it’s quite probable that the corporation is use an advanced version of Apple Silicon instead of Google Tensor for model training.
Whatever the case, this does indicate that Google has played a part in the iPhone’s increased intelligence through Apple Intelligence.