Latest Business News
Qualcomm and Meta Team Up to Bring Llama 2 to Mobile Devices
Qualcomm and Meta Collaboration
Qualcomm and Meta have announced their partnership to enable the new large language model, Llama 2, to run on Qualcomm chips on phones and PCs. This collaboration will begin in 2024, marking a significant shift for large language models, which have primarily run on server farms using Nvidia graphics processors. Qualcomm aims to position its processors as ideal for artificial intelligence (AI) applications on the edge, or on the device itself, rather than relying on cloud-based computing.
Advancements in AI Technology
Running large language models on mobile devices can potentially lower the cost of running AI models and result in improved and faster voice assistants and other applications. Qualcomm's tensor processor unit (TPU) is designed for the complex calculations required by AI models. However, the processing power available on mobile devices is far less compared to data centers equipped with advanced GPUs.
Meta's Llama 2 on Qualcomm Devices
As part of the collaboration, Qualcomm will make Meta's open-source Llama 2 models available on its devices. This collaboration will enable applications such as intelligent virtual assistants. Llama 2 offers similar capabilities to ChatGPT but in a more compact program that can run on mobile phones.
The Significance of Meta's Open-Source Llama 2
Meta's decision to publish the "weights" of Llama 2 sets it apart from other notable large language models. This move allows researchers and commercial enterprises to use the AI models on their own computers without seeking permission or paying. Unlike closed-source models such as OpenAI's GPT-4 or Google's Bard, the weights of Llama 2 are openly available.
Prior Collaboration and Demo
Qualcomm has previously collaborated with Meta on chips for the Quest virtual reality devices. The company has also demonstrated the capabilities of AI models running on its chips, including the open-source image generator Stable Diffusion. This history of collaboration highlights the ongoing partnership between Qualcomm and Meta in advancing AI technology.
Revolutionizing AI on Mobile Devices
The collaboration between Qualcomm and Meta to bring Llama 2 to mobile devices marks a significant milestone in the field of AI technology. With Qualcomm's powerful tensor processor unit (TPU) and Meta's open-source Llama 2 models, the potential for enhanced AI applications on smartphones and PCs is immense.
One of the key takeaways from this collaboration is the shift towards running large language models on edge devices rather than relying on cloud-based computing. This opens up new possibilities for businesses looking to leverage AI technology without the need for expensive server farms. With Llama 2 running on Qualcomm chips, the cost of running AI models could be significantly reduced, enabling small and medium-sized businesses to harness the power of AI more affordably.
Additionally, the decision by Meta to publish the weights of Llama 2 as open-source distinguishes it from other large language models in the market. This move democratizes access to the AI model, allowing researchers and businesses to utilize it without any restrictions. This could foster innovation and collaboration, leading to the development of new and exciting applications that leverage Llama 2's capabilities.
The prior collaboration between Qualcomm and Meta, as seen with their work on Quest virtual reality devices, establishes a strong foundation for further advancements in AI technology. As Qualcomm continues to demonstrate its prowess in running AI models on its chips, businesses can expect even greater opportunities for leveraging AI to drive growth and innovation.
In conclusion, the Qualcomm and Meta partnership, bringing Llama 2 to mobile devices, holds immense potential for transforming the AI landscape. Businesses that embrace this technology and explore its possibilities stand to gain a competitive edge in an increasingly AI-driven world.
Article First Published at: https://www.cnbc.com/2023/07/18/meta-and-qualcomm-team-up-to-run-big-ai-models-on-phones.html