The release of DeepSeek, which was reportedly trained at a fraction of the cost of leading models, has solidified open-source AI as a serious challenge
The release of DeepSeek, which was reportedly trained at a fraction of the cost of leading models, has opened up new possibilities for open-source AI, according to Dr. Ala Shaabana, co-founder of the OpenTensor Foundation.
In an interview with Cointelegraph, Dr. Shaabana discussed the recent advances in open-source AI and how they compare to centralized projects. He also shared his thoughts on the future of AI and the role that open-source will play in its development.
According to Dr. Shaabana, the release of DeepSeek, an open-source AI model that was reportedly trained at a fraction of the cost of leading models, hassolidified open-source AI as a serious challenger to centrally managed projects.
“The entire paradigm of centrally managed AI, which costs tens of billions of dollars to develop and train, is being called into question by DeepSeek,” the OpenTensor Foundation co-founder told Cointelegraph. “Open-source AI is now capable of achieving performance that is on par with — and in some cases, even exceeds — the best centralized models.”
Dr. Shaabana attributed the rapid progress of open-source AI, and the narrowing of the gap between centralized systems, to a procedural shift in academia, which is now requiring researchers to include their code with their papers in order to submit to academic journals for publication. This requirement, he said, is making it easier for other researchers to build upon existing work and to develop new models more quickly.
“Another factor that is contributing to the success of open-source AI is the increasing regulatory burden on centralized AI projects,” Dr. Shaabana added. “For example, there are now potential geographic restrictions on data due to geopolitical tensions, which could place further burdens on centralized AI projects.”
The costs and regulatory burdens resulting from increased regulations would widen the performance gap between centralized and open-source systems, which are not susceptible to those constraints, he explained.
“We are already seeing the effects of this dynamic in the market, with several open-source AI startups attracting large investments and centralized AI projects struggling to keep up,” Dr. Shaabana said. “I believe that this trend will continue in the future, and that open-source AI will play an increasingly important role in the development of AI.”
