## The Rise of Open-Source AI
Open-source AI is experiencing a surge in popularity, with Meta’s Llama family of models leading the way. According to Meta’s mid-year update, the downloads for Llama models on Hugging Face have increased ten-fold from a year ago, approaching 350 million. This growth is attributed to the adoption of Llama models by large enterprises such as Zoom, Spotify, Infosys, AT&T, and Goldman Sachs, who are utilizing them for both internal and external use cases.
This shift towards open-source AI is challenging the dominance of closed-source models and raising questions about the lead of OpenAI. While OpenAI had an early advantage in the generative AI field, Meta quickly caught up with its Llama model. Unlike OpenAI, which focused on closed-source models, Meta embraced an open approach, making Llama models available through Hugging Face and cloud partners.
## The Benefits of an Open Strategy
Meta’s open strategy has resulted in a vibrant and diverse AI ecosystem, where developers have more choice and capability than ever before. This approach has also led to the creation of powerful open models by other companies, such as French startup Mistral. Developers can build upon and create derivatives of these models, achieving parity or even outperforming closed models on select metrics.
The success of Meta’s open strategy can be seen in the growing preference for Llama models within the developer community. According to a survey by Artificial Analysis, Llama was the second most considered model and the industry leader in open source. Enterprises like AT&T, DoorDash, Goldman Sachs, Niantic, Nomura, Shopify, Spotify, Zoom, Infosys, and KPMG are among those utilizing Llama models for various use cases.
## Growing Adoption and Pressure on Closed Models
The adoption of Llama models extends beyond Hugging Face, as many enterprise developers are leveraging them through Meta’s network of cloud and infrastructure providers. While specific partner numbers were not disclosed, Meta confirmed that Llama’s monthly usage grew ten-fold from January to July 2024 for select cloud service providers.
This growing adoption of open-source AI, particularly by leading enterprises, puts pressure on companies offering closed models. The recent improvements in performance and long-term cost benefits of the open approach are driving this shift. As open-source AI continues to develop, it is likely to challenge the domination of closed models and push companies to innovate further and reduce the cost of using their models.
## OpenAI’s Challenges
While OpenAI has made significant price reductions for its existing models, including GPT-4o, it appears to be lagging in terms of product innovation. The research lab, led by Sam Altman, has announced cutting-edge AI products such as Sora and SearchGPT, but they remain unreleased or limited to a select group of users. This lack of product innovation raises questions about OpenAI’s ability to keep up with the advancements in open-source AI.
In conclusion, open-source AI, led by Meta’s Llama models, is gaining traction and challenging the dominance of closed-source models. The adoption by major enterprises and the benefits of the open approach are driving this shift. As open-source AI continues to evolve, it will put pressure on closed models and encourage further innovation in the field.