Kai-Fu Lee, a renowned AI entrepreneur, is shifting his startup's focus towards DeepSeek and open-source models. This strategic move is driven by the rising costs of training large models, which has become a significant barrier for startups.
Lee's startup, (link unavailable), has realized that pre-training massive models is no longer cost-effective for them. Instead, they're focusing on developing smaller, industry-specific models that can deliver high commercial value. This approach allows them to prioritize faster, smaller, and more cost-effective models.
The partnership with Alibaba Cloud is also a crucial aspect of this strategy. By collaborating with a tech giant, (link unavailable) can leverage Alibaba's capabilities in training large models, while focusing on developing smaller models that can be applied to various industries.
Lee emphasizes the importance of adapting to the changing AI landscape. He believes that the industry is moving towards a more application-focused approach, rather than solely pursuing scaling laws. This shift is driven by the need for more cost-effective and practical AI solutions.
In terms of open-source models, Lee sees an opportunity for (link unavailable) to contribute to the development of more accessible and affordable AI technology. By embracing open-source models, the startup can reduce costs and increase efficiency, ultimately driving innovation in the AI industry.
