A policy retreat in the US artificial intelligence space has driven interest in open-source models, particularly those from China. This shift challenges the traditional model of closed models and points to a more decentralized future.
In the context of growing skepticism about government regulation, we examine the AI industry's response to the regulatory clampdown in the US. Under the leadership of an industry-friendly president, the US government has taken unusual measures to restrict access to some of the most advanced AI systems developed by Anthropic and OpenAI. Despite this, interest in open-source models has surged, particularly those from China. This trend challenges the traditional model of closed models, which had dominated the industry until now, and points to a more decentralized future for AI.
It's time to question the convenience of closed models, which allow companies to control access and availability. Allowing the industry to continue advancing in this way, subject to the discretion of authorities, can have serious consequences. The lack of standardization and interoperability becomes a significant problem, as applications dependent on these models become less reliable when faced with unexpected obstacles, such as government-imposed access blocks.
The experience of Anthropic and OpenAI serves as a clear example of the consequences of relying on closed models. Startup Anthropic, in response to the government's order to restrict access to its most advanced models, simply disabled them entirely, leaving its users without a solution. On the other hand, OpenAI opted to let the authorities approve each of its clients, which, according to analysts, does not solve the problem of the industry's fragility.
The impact of this situation is evident in the frustrated experience of developers of AI-dependent applications. The lack of stability and volatility of these models forces them to reconsider their strategies and seek more resilient alternatives. In this context, open-source models offer an alternative vision to the traditional AI business model, based on closed and regulated models. In the future, the AI industry will have to navigate a more complex and decentralized environment. Some analysts anticipate that this boom in open-source models could drive greater collaboration and sharing of resources, allowing for improved innovation and competitiveness in the sector. While the transition to a more decentralized model carries risks, it will also offer opportunities to develop new, more reliable, and more accessible solutions.