Apple in talks with startup that shrinks AI models to run on an iPhone
PrismML says its compressed models use up to 15 times less memory and run 6 to 8 times faster, a pitch Apple is evaluating for Siri.
- Apple is evaluating technology from Silicon Valley startup PrismML that could allow powerful AI models to run directly on iPhones, with CEO Babak Hassibi telling CNBC that "things are progressing nicely."
- Most capable AI models require significant memory and processing power, creating a central constraint for Apple, which PrismML aims to address by reducing model size from roughly 54 GB to less than 4 GB.
- PrismML compresses models by reducing values from 16 bits to one or three, using between 10 and 15 times less memory, a method Hassibi compared to the chip industry's shift from eight-bit to four-bit computing.
- Running AI locally could help Apple improve Siri without sacrificing privacy, as Horace Dediu of Asymco noted Apple is likely trying to fit more capable models within physical device limits to reduce latency and cloud costs.
- Analysts caution that PrismML's claims require proof outside controlled demonstrations, as Tarun Pathak of Counterpoint Research stated the "ultimate test will be millions of queries," with battery consumption remaining a critical open question.
20 Articles
20 Articles
Apple is reportedly in talks with a startup that possesses technology capable of reducing the memory usage of large artificial intelligence (AI) models by up to one-fifteenth. If the technology is successfully verified, Apple's on-device strategy of processing Siri's AI computations within the iPhone rather than in the cloud is expected to be further strengthened. According to CNBC on the 14th (local time), PrismML, a spin-off from the Californi…
Apple looks to shrink AI models for iPhones
Apple is reportedly in very early discussions with PrismML, a Silicon Valley startup spun out of the California Institute of Technology and backed by Khosla Ventures, about technology that could dramatically reduce the size of powerful AI models so they can run directly on iPhones, CNBC reports. PrismML recently announced that it compressed Alibaba’s open-source 27 billion parameter Qwen model from roughly 54 GB to less than 4 GB, allowing all 2…
PrismML confirms it is in talks with Apple about AI model-shrinking tech
Five days after initial reports that there was an AI-model shrinking technology by PrismML suitable for iPhones, the company has confirmed it is talking to Apple over the use of its technology.Siri at work on an iPhoneOne of the major problems with AI processing is the need to manage massive models that most normal computers and smartphones cannot easily handle alone. On July 9, the start-up PrismML surfaced with a solution to shrink down the mo…
PrismML releases Bonsai 27B, claiming first major AI model of its size fit for iPhone
PrismML is back in the news today after the AI startup’s CEO told CNBC that Apple is looking at the company’s technology. The company has also released its Bonsai 27B model that it says runs on iPhone, iPad, and Mac.
Coverage Details
Bias Distribution
- 67% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
