The world is dominated by variations of AI. And the notions of developing such models for pursuit of expansive AI require much work in terms of technical resources and not to mention, financial investments. Hence, the US government had sworn to support the Stargate Project which was announced by President Donald Trump. To their dismay, the AI development firm DeepSeek in China has played havoc with the US’s plan set in motion.
With the introduction of DeepSeek R1LLM (Large Language Models) it presented “a state-of-the-art, open-source reasoning model and a detailed recipe for training such models using large-scale reinforcement learning techniques”.
DeepSeek is also providing its R1 models under an open source license implying its free use. China’s deepseek into this AI model has revealed incapability in the US’s longstanding victory in innovation and technology. The R1 model was developed for less than 6 million dollars according to its claim, which offered a surprise attack to the US tech companies that had invested billions in AI.
DeepSeek’s capabilities have outshined the reliance that the US expected on its technology as DeepSeek portrayed the development of AI to new heights despite the restriction of export of the “highest-performance AI accelerator and GPU chips from the US”. Marc Andreessen (tech investor) “called it “AI’s Sputnik moment”, comparing it to the Soviet Union’s space race breakthrough in the 1950s.”
In the research paper by DeepSeek titled “DeepSeek- R1: Incentivizing Reasoning Capability in Large Language Models via Reinforcement Learning” offers the basis of the self-evolution process of the model. “Through reinforcement learning, the model naturally learns to allocate more thinking time when solving reasoning tasks. Amazingly, this occurs without any external adjustments.”
Coming to another phenomenon mentioned in the research paper, the “aha moment” of DeepSeek -R1 Zero, “the model learns to revaluate its initial approach and correct itself if needed”.
This isn’t the only arena in which China has or is surpassing the US. In terms of renewable energy, “China dominates solar power production and leads global investments in wind power, while the US remains mired in political debates over green energy”. Not to mention the conundrum arising from executive orders signed by President Trump after he was sworn in.
The increase in costs of extensive research and development/advancements with over dependency on the private sector on the private sector and investments made by private companies have ridden the US tirelessly, keeping it at a slower pace as well as a state of dysfunction.
According to the Times of India, today, DeepSeek ban has begun in the US, with Texas as the first state to take initiative to do so. Greg Abott, the governor of Texas stated, “Texas will not allow the Chinese Communist Party’s to infiltrate our state’s critical infrastructure through data harvesting AI and social media apps”. The US Navy has also blocked access due to national security concerns.
The continuous capacity of AI with argued abilities such as the method wherein an AI system learns from another AI system would arguably make stopping DeepSeek’s ability to acquire Intel and knowledge from AI in the US such as Open AI more difficult. This method is referred to as “distillation”. The US appears to be banking on this argument, although, as already aforementioned, this method is difficult to stop.