Qualcomm, a giant in the field of high-end smartphone processors, is gearing up to unveil its third-generation Snapdragon 8 chips. The announcement is especially noteworthy considering that its current top-of-the-line predecessor, the Snapdragon 8 Gen 2, currently holds the title of the second most powerful smartphone chip of all time, according to . Antutu Benchmarkonly the best is Dimensity 9200+ developed by Taiwanese manufacturer MediaTek.
Details of new chip to be officially announced on Tuesday leaked to the internet on monday. The new chip promises not only improved performance but also a significant leap forward in on-device AI capabilities. MS power user.
In addition to expected incremental performance improvements, such as 25% faster GPUs and 30% faster CPUs, new systems-on-chips (SoCs) have powerful features that can support high accuracy for local AI models. It is equipped with an AI engine. Run local large-scale language models (LLMs) with over 10 billion parameters. The chip can also perform stable diffusion of text-to-image models using the ControlNet neural network.
Image: MSPoweruser
Mobile hardware manufacturers are making a strong pivot toward AI. For example, Google has improved its Tensor chip to allow its latest Pixel smartphones to manage the myriad AI functions it performs locally, bypassing the need for cloud processing. Similarly, Apple has upped its game with his A17 Pro SoC and enhanced its AI capabilities.
The advent of AI-enabled chips in consumer devices heralds an era of technological change. These advanced processors are designed to run complex AI models locally, from personalized chatbot assistants and hyper-realistic gaming experiences to real-time health monitoring and adaptive user interfaces. It suggests a future full of possibilities.
Promising faster processing and greater personalization and increased privacy, these chips will ensure that our devices are not just smart, but intuitively adapted to our needs and preferences. It is positioned as the foundation of the future.
Kent Carsay is CEO. call AIis a company at the forefront of AI image generation. Harnessing the power of stable diffusion.he said Decryption There are many benefits to running AI on your device rather than in the cloud.
“Cloud-based computing is a great option for large enterprises, but for individuals looking to create content, on-device AI offers a more affordable, personalized and private alternative,” he said. ” he said. He also highlighted the potential of Apple’s approach in introducing large-scale unified memory, and suggested, for example, a future of using a powerful local LLM assistant instead of cloud-based Siri.
But while the new chip is powerful, Carsay said Invoke has no plans to port the app to Android or iOS devices.
“Our commercial solutions are typically used for more complex workflow processes,” he said. “While we have no immediate plans for a local His OSS Android app, we are closely monitoring the trajectory of creative usage patterns and the emergence of cutting-edge technologies to support mobile deployment.”
Currently, the go-to option for local image generation is the Draw Things app on iOS. There is no comparable option for Android users.
Carthay was optimistic about the ripple effect of Qualcomm’s new chips in the broader technology ecosystem.
“Providing this chip for on-device AI increases the capabilities that can be brought to mobile devices,” he explained, adding that as hardware capabilities improve, app developers will be able to service on-demand AI capabilities. He added that he expects it to be incorporated into the
However, with every innovation there are challenges. Kersei pointed out potential problems in distributing model weights to end devices, especially for proprietary models like GPT-3 and GPT-4, which are large and not open source.
“In a proprietary model where managing intellectual property is paramount, this may not be viable,” he said. But Kersei was quick to note that open source models like Stable Diffusion and Mistral7B could be a game-changer.
According to reports DecryptionMistral7B is a lightweight model trained with 7 billion parameters that has become a hot topic in the AI community. Decryption tested the response against other strong competitors such as LlaMA and Stable Beluga 2 and outperformed their output. By the way, Mistral 7B operates with his 7 billion parameters, while GPT-4 boasts his staggering 1.7 trillion parameters.
Qualcomm’s Snapdragon 8 Gen 3 is more than just a microchip. It’s a harbinger of an AI-centric mobile future. With industry leaders like Qualcomm, Google, and Apple leading the way, it’s only a matter of time before Assistant can actually help you answer questions and perform tasks while keeping all your secrets safe. .