Chinese AI startup DeepSeek announced it will begin publicly releasing key code and data starting next week, Bloomberg reports.
This bold step towards open-source AI development surpasses the level of transparency currently practiced by industry giants like OpenAI and Meta, potentially accelerating the global AI race.
DeepSeek, a relatively young company at just 20 months old, has already garnered significant attention for the sophistication of its AI models, outperforming competitors in recent benchmark tests. Now, the company is taking its challenge further by making its code repositories freely available to all developers and researchers.
This unprecedented level of transparency will allow anyone to download, study, and build upon the code that powers DeepSeek’s R1 and other platforms. In a post on X, the company stated its commitment to sharing its progress “with full transparency.”
“We’re a tiny team exploring AGI. Starting next week, we’ll be open-sourcing 5 repos, sharing our small but sincere progress with full transparency,” the company announced on its X handle on Friday.
DeepSeek’s decision represents a significant push towards a more open-source approach to AI development, a philosophy gaining momentum as its models demonstrate superior performance compared to those from OpenAI and Meta. While companies like Meta already release their models to the public, allowing for customization, DeepSeek aims to go a step further by publicizing the underlying code, the data used to train its models, and the methodologies employed in development and management.
This move also has geopolitical implications, potentially escalating the competition between the US and China in the race to develop increasingly advanced AI models. By open-sourcing its coding secrets, DeepSeek hopes to foster wider adoption of its technology.
However, the move also raises concerns about security. Governments from the US to Australia are already grappling with the potential risks associated with the widespread availability of sophisticated AI, and DeepSeek’s decision is likely to amplify these concerns.
