Skip to content

The Battle Over AI Dominance: Proprietary vs. Open-Source Models

Large language models (LLMs) are at the center of a clash between proprietary and open-source models in the field of AI. Giants like Microsoft and Google are defending their proprietary technology against open-source LLMs, as the industry grapples with potential commoditization of AI capabilities.

The ongoing tussle for control over pivotal technology has been a constant in the tech sector's history. Unsurprisingly, this struggle has now seeped into the realm of artificial intelligence (AI), where the battleground involves the clash between free open-source systems and patented generative AI products like ChatGPT.

At the core of this conflict lie large language models (LLMs), intricate algorithms that form the heart of AI. The AI boom has led to a face-off between industry giants as Big Tech firms safeguard their proprietary technology against contenders drawn to unfettered programs. Advocates of open-source AI assert that it will democratize access to AI tools, making them more affordable and facilitating new LLM development and commercialization. However, Wall Street analysts covering AI stocks like Microsoft and Alphabet express concerns that open-source AI could turn proprietary models into commodities.

Microsoft, a major investor in OpenAI, stands as a prime example. Despite the moniker, OpenAI operates proprietary systems such as ChatGPT, which are perceived as more advanced than many rivals. Google is also working on a proprietary LLM called Gemini. However, companies like Meta Platforms (Meta) and Amazon are championing the open-source cause. Meta's open-source LLMs are advancing swiftly due to their availability to researchers and developers, potentially benefiting content creators, small businesses, and advertisers.

While Microsoft and Google hold considerable sway, open-source AI models from other tech giants are gaining ground, thanks in part to Meta's initiative. Amazon is collaborating with multiple open-source LLM developers, suggesting that if LLMs become commoditized, Amazon could find success by providing alternatives. With open-source LLMs, businesses might favor constructing apps rather than licensing proprietary technology.

Notably, open-source LLMs are becoming more efficient, requiring less computing power for training. These models are revolutionizing human-AI interaction, making it easier for users to engage with AI systems without algorithmic understanding. However, the crux of generative AI lies in large data sets called parameters used to train LLMs. The size of these parameters determines the LLM's capabilities, with models like OpenAI's GPT-3.5 trained on 175 billion parameters.

Amid the race for AI supremacy, competitive moats, or barriers to entry, are critical for Wall Street. While various large language models possess distinct capabilities, OpenAI remains ahead in the LLM arena. This performance edge would affect Microsoft's return on its substantial investment in OpenAI.

Meanwhile, the open-source debate resonates with the history of technology, including the evolution of open-source programs like Linux. While competitive moats are preferred on Wall Street, the open-source AI landscape is anticipated to coexist with proprietary models. The implications of this emerging market for generative AI are manifold, with startups potentially challenging tech giants and reshaping the industry.

Ultimately, the balance between open-source and proprietary models will continue to evolve, potentially shaping the future of AI dominance.