DeepSeek's Vision: Driving Core AI Innovation, Challenging China's Follower Mentality, and Building an AGI Ecosystem

This interview with DeepSeek CEO Mr. Liang Wong delves into the company's disruptive entry into the global AI market, characterized by its surprisingly low pricing for large language models. Wong clarifies that DeepSeek's intention was not to instigate a price war but to make AI technology affordable and accessible to everyone, driven by their ability to significantly lower costs through advanced research. He emphasizes that DeepSeek's ultimate goal is the pursuit of Artificial General Intelligence (AGI), which necessitates fundamental research into new model structures and data efficiency, rather than merely building applications on existing models like Llama, which he notes are generations behind leading Western models.

A central theme of the discussion is China's historical role as a technological follower, often borrowing ideas and prioritizing quick market entry over original innovation. Wong argues that China's current lag is not due to a lack of funding—given the substantial profits of its tech giants—but rather a deficit in self-confidence, experience in talent integration, and a creative, curiosity-driven attitude. He asserts that Chinese companies, including DeepSeek, must shift from a "reverse engineering" mentality to one of "original innovation" to truly contribute to global technological advancement and avoid perpetually playing catch-up, especially in critical sectors like AI.

DeepSeek's commitment to open source is presented as a strategic choice, not a vulnerability. Wong explains that in an era where even closed-source models are quickly caught up to, the true barrier to entry lies in building an innovative organization, fostering a strong team culture, and accumulating know-how. Open sourcing their advancements serves as a badge of honor, attracting top talent drawn to solving difficult problems and contributing to a broader technological ecosystem. This approach contrasts with the traditional protectionism and quick-profit motives that have historically limited Chinese innovation.

Looking ahead, Wong is optimistic about the future of AI and China's role. He believes that the global economy's structural transformation will increasingly rely on core innovation, pushing Chinese society to re-evaluate its priorities away from "quick money grabbing." DeepSeek's roadmap for AGI focuses on mathematical/coding capabilities, multimodal interactions, and natural language processing, with an estimated timeline of 2-10 years for achieving true AGI. He envisions DeepSeek leading in LLM research while partners build applications, fostering an ecosystem that positions China at the technological frontier, despite challenges like geopolitical sanctions impacting access to high-end AI chips.

It was not our original intention to become a game changer or Market disruptor it was by accident.

our principle is not to lose money but also avoid overcharging our customers.

AI or AGI should be a technology that is Affordable to everyone.

our goal is Agi this means that we need to explore new model structures... we consider this process to be basic research needed in order to scale up to larger models.

China basically didn't participate in any real technological innovations we're just too used to wait for someone else to come out with better Hardware better software.

what we are lagging is certainly not funding is the lack of self-confidence and the lack of experience in putting together talents to achieve these Innovations.

for this kind of break through advancement the time you can buy from protectionism is limited.

being open source is a display of attitude is not profit driven no we wear this achievement as a batch of Honor we believe this will attract more talents.

the real difference is original Innovation uh versus let's say reverse engineering if we cannot change this mentality as a follower we will always be the one who's playing catchup.

the problem we Face were never funding but the high-end AI chips uh that is not available to us due to sanctions.

I think AI is at a stage where there will be many technological bigru I think the real error for AI application has not yet arrived.

there's no top Super Genius just fresh graduates from top universities uh PhD students and young talents with a few years of experience under that belt.

I think to innovate is to believe.

the brightest Talent are often attracted to the most difficult task uh to solve the most difficult questions in the world.

maybe 2 years 5 years 10 years is should be within our lifetime.