WASHINGTON, March 31 (QNA) Google has said that random memory is increasingly being used to power artificial intelligence models in smartphones.
The American company had earlier announced its intention to offer the Gemini Nano artificial intelligence model on its Pixel 8 Pro phone, while the other version, the Pixel 8, will not get the model due to restrictions regarding hardware specifications.
Later, Google announced that this model would be available on the Pixel 8, but as a “developer option,” meaning users could only enable it through the hidden Developer Options menu in Settings. Model on Pixel 8 Pro phone by default in user interface.
Xiang Zhao, Google's vice president of devices and services programs, explained the company's decision through an official audio program (podcast) associated with Google, noting that the Pixel 8 Pro phone has 12GB of random memory (RAM). Suitable for running Gemini Nano model. The Pixel 8 has less than 8GB of RAM, and the model is not easy to operate.
According to Zhao, Google's fear is that the company doesn't want to “diminish the experience” on the Pixel 8 phone with less RAM compared to the other version.
Zhao added that the company wants some AI-backed features like Smart Reply to be in “random memory,” meaning the model will permanently occupy a portion of random memory in preparation for activating those features at any time.
On the other hand, recent leaks indicate that Apple is working on increasing the random memory and storage capacity in the base versions of the upcoming iPhone 16 phones in preparation for adding artificial intelligence models capable of working on these devices.
