Chinese AI model challenges industry leaders with more efficient and low cost technology

Chinese startup Deepseek is drawing attention to the artificial intelligence community with the launch of the DeepSeek-R1an open source model that quickly surpassed OpenAi’s chatgpt on Apple’s free list of applications. Developed with a much smaller budget than its competitors, the model promises high performance in complex tasks such as mathematics, coding and reasoning in natural language.

Different from traditional approaches, the Deepseek-R1 adopts the Reinforcement Learning (RL) As a central training method, allowing the model to evolve through interaction and rewards, becoming more efficient and adaptable. This strategy challenges the dominant tendency of the sector, which has been investing in the increase in computational power for years to train IA.

Experts point out that Deepseek’s success can change the course of the industry. Economist Jeffrey Sachs pointed out that the model proves that it is possible to develop was advanced at much lower costs than imagined. Investor Marc Andreessen has described Deepseek-R1 as “one of the most amazing advances” has ever seen, while Banco Morgan Stanley considers the startup’s approach a viable alternative to “arms race” for computational power.

With the motto “Greater is no longer intelligent”, DeepSeek emerges as a strong name in the AI ​​revolution, proving that innovation and efficiency can redefine the future of the industry.