Whip bulls reported that on February 26, according to foreign news reports, Google is optimistic about the prospect of using more advanced artificial intelligence models on smartphones next year.
The internet giant expects its currently available Gemini large language model (LLM) to compete with OpenAI's most powerful GPT-4 AI model powered by Microsoft, which will begin to be embedded in devices starting next year.
Google has already made the Gemini Nano available on its Pixel devices and all other capable Android devices, which is the company's most effective on-device AI model.
Brian Rakowski, vice president of product management for Google's pixel division, said he expects the company's state-of-the-art large language models, which are currently only accessible through remote data centers via an internet connection, to land directly on smartphones next year.
There is a smaller version of our Gemini model in the cloud, Rakowski told CNBC. Quite a few breakthroughs have been made in compressing these models so that they run on devices. 」
Some have already been validated, and some are being explored for certain applications. It would be great if you had all the models on the device. It still has fantastic applications. 」
The Gemini Nano is at a level of performance comparable to that of our **model less than a year ago,adds Rakowski. You can do a lot of things on your device with these smaller, refined models. 」
If you follow this trajectory, we think that some of the things that we have to move to the cloud next year are going to be on devices, which is very exciting, and it's instantaneous and doesn't require a connection or subscription. 」
Large language models (LLMs) are AI models that are capable of understanding and generating language in a human-like manner. Gemini Ultra is Google's top LLM with a high number of parameters of 156 trillion. In comparison, OpenAI's GPT-4 is powered by 176 trillion parameters.
After several years of experiencing a massive slowdown in device sales, smartphone makers have been dreaming of a supercycle in their industry, fueled by artificial intelligence. In 2023, smartphone sales drop to 11600 million units, the lowest number of shipments in a decade.
Analysts say a supercycle is unlikely to emerge in the next few years, as there aren't enough novel features and innovations on the market to convince people to upgrade with aging smartphones.
Francesco Jeronimo, vice president of data and analytics at research firm IDC, told CNBC: Unfortunately, we don't expect this kind of boom.
The last supercycle we saw was between 2010 and 2015, when the market grew fivefold in five years, from about 300 million smartphones per year to 1.5 billion. 」
Still, more and more smartphone manufacturers are investing heavily in AI in the hope that it will spark more interest in mobile technology.
Companies like Humane, Rabbit, and China's Meizu are betting on the future of smartphones that don't even look like traditional smartphones. These devices will be smaller and more compact, and we can interact with them through voice activation, just like the Amazon Echo speakers, but on the go.
Google has been investing heavily in artificial intelligence in hopes of gaining an edge over competitors like OpenAI (the Microsoft support company behind ChatGPT).
Google recently announced a major rebranding of its ChatGPT replacement, Bard, including brand new apps and subscription options. Bard was renamed Gemini, which bears the same name as the suite of AI models that power chatbots.
Android users can ** Gemini dedicated Android app, while iPhone users can use Gemini in Google app on iOS.
Alphabet CEO Sundar Pichai highlighted the company's commitment to artificial intelligence during the company's January 30 earnings call**. Pichai said he ultimately hopes to provide an artificial intelligence that can complete more and more tasks on behalf of users, including in Google searches, although he said there is still a lot of work to be done.
Similarly, CEOs of tech giants from Microsoft to Amazon have emphasized their commitment to making artificial intelligence** a productivity tool.