😶‍🌫️The Washington Post recently pointed out one interesting fact about the huge change in AI.

The Washington Post recently pointed out one interesting fact about the huge change in AI.

They say that Chinese big tech companies are taking the lead in the ecosystem, beating the U.S. by using the Open-Weight AI model.

Chinese companies took the opposite path while Meta and OpenAI refined their top-performing models behind closed doors. It made the model unwound to the world, made it available to anyone, and in the process, innovation spread.

At this point, China’s choice appears to be more than just a pursuit strategy. They are taking a sophisticated strategic balance in which they take the initiative in application and derivation while disclosing their core technologies.

Developers around the world create derivatives from them when the model is solved, which then seeps back into the industrial field. In the end, the entire ecosystem is built on their standards.

Many professors and corporate research institutes around me have already installed a lot of them in their labs to experiment, saying that the performance of the Chinese model is quite good and that it doesn’t matter if it’s from China. As such, the spread of the open ecosystem is fast, and at this rate, not only Korea but also the whole world may use AI under the rules made by China.

Korea is also creating performance models through its own FM (foundation model) business, and some of them will be released in the form of open sources that schools, research institutes, and companies can freely use for research and commercial purposes starting this December.

If the plan goes smoothly, Korea will be able to establish itself as an open innovation hub, not just a technology importer.

Based on public data and the collective intelligence of research institutes led by the state and Korean AI companies, I hope there will be a foundation for Korean researchers and startups to experiment to their heart’s content.

Seeing that LLM models are now being trained to some extent, the competitiveness of the new era will be who can use it with more people than who has a stronger model.

We hope that the emergence of open-source basic models prepared by Korea, which will be released in December this year, will be the first step toward its opening.

출처: https://www.washingtonpost.com/technology/2025/10/13/china-us-open-source-ai/