Author: Techub News Compilation

Introduction

In June 2025, Alexandr Wang, founder and CEO of Scale AI, appeared on the well-known podcast "Shawn Ryan Show" for an in-depth conversation lasting 204 minutes. As a leading figure in the AI data infrastructure field and one of Silicon Valley's youngest billionaires, Wang's insights profoundly influence the industry's thoughts on the development path and security of AI. In this interview, he not only explored the future landscape and potential risks of brain-computer interfaces (such as Neuralink) but also analyzed the fierce competition between the U.S. and China in AI, chips, and data from a national strategic perspective, emphasizing the extreme importance of maintaining "human sovereignty" amidst the rapid enhancement of AI capabilities.

Summary

• Brain-computer interfaces are a bridge to "superhuman" cognitive abilities, but they also represent an unprecedented vector for attacks that could be used to manipulate perception, emotions, and even memory.
• Data is the "new oil" of the AI era, while computational power (chips and data centers) is the key bottleneck that determines AI capabilities, directly affecting future military and economic advantages.
• The U.S.-China AI competition is a "war" concerning future global dominance; the U.S. must maintain comprehensive leadership in algorithms, data, and computing power to form effective deterrence.
• The "recursive self-improvement" of AI capabilities could lead to a dramatic widening of the gap between leaders and followers, forming a winner-takes-all situation.
• AI systems must be constructed with "human oversight" at their core to ensure humans maintain ultimate control over critical decisions, which is fundamental to preventing catastrophic consequences.

Brain-Computer Interfaces: The Allure of Superpowers and Ultimate Risks

The interview began with discussions on brain-computer interface technologies like Neuralink. Alexandr Wang adopted a pragmatic and forward-looking attitude toward this topic. He believes that as AI capabilities exponentially outpace the slow evolution of human biology, humans will ultimately need to directly connect with AI to tap into its powerful cognitive abilities and avoid being left behind by the times. He envisions that future humans using brain-computer interfaces will possess encyclopedia-like knowledge, ultra-fast thinking, and the ability to process global information like AI, thereby transforming into "superhumans" cognitively.

However, Wang does not shy away from acknowledging the enormous risks involved. Directly connecting the brain to computers and the internet creates an unprecedented attack surface. He paints a terrifying scenario: companies might implant advertisements directly into your thoughts; more dangerously, foreign actors, terrorists, or hostile regimes could invade the brain, stealing memories, controlling thoughts, or engaging in manipulation. When technology advances to the point where sensory information (visual, auditory, tactile) can be precisely projected or even emotions directly injected, shaping a completely false reality will become possible. Wang admits that this level of "propaganda manipulation" will leap from the current low intensity of 1-2 to a high intensity of 9-10, greatly enhancing control over individuals and society.

When asked if he would consider implantation for himself, Wang answered affirmatively but added strict conditions: first, there must be absolute confidence in the cybersecurity defenses of brain-computer interfaces; second, it needs to be assured from a large amount of pioneer data that this technology will not fundamentally and uncontrollably alter consciousness. He further speculated that brain-computer interfaces could be the elementary stage of "uploading" human consciousness to computers, thereby achieving some form of "digital immortality."

Data as Oil, Computing Power as the Engine: The Equation of National Power in the AI Era

As the helm of Scale AI, Wang positions the company as the "backbone" of the AI revolution, providing the data and infrastructure required for large enterprises and governments to deploy secure, advanced AI systems. He likens data to the "new oil" of the AI era, serving as the core fuel for driving intelligent algorithms. The other two major pillars of building AI capabilities are algorithms (software) and computational power (hardware).

He places particular emphasis on the strategic importance of computational power, specifically surrounding chips and data centers. Future AI competition isn't just about having the most advanced models; it's about scale: how many powerful AI instances can you run simultaneously? This directly depends on how many data centers you can build and supply, how much power you can obtain, and how many chips you have. Wang points out that a key front in the competition among giants like OpenAI, xAI, Google, and Meta is the race to secure electricity and data center construction sites years in advance, as the advantage of the future is already being laid out today.

This logic brings geopolitical risks to the forefront. Wang notes that about 95% of high-end chip manufacturing is concentrated in Taiwan (TSMC). If mainland China were to control Taiwan through blockade or invasion, it would seize the most critical strategic resources in the AI competition. He analyzed the dilemmas of various response options: Taiwan itself may be unwilling to destroy these wafers that are vital for its survival; U.S. bombing could directly escalate to a world war. Therefore, the best course is to deter conflict through powerful deterrents (including military and AI capability advantages). However, this also means that the U.S. must accelerate its domestic chip manufacturing capabilities, despite facing multiple challenges such as regulations, power, talent, and TSMC's own strategic considerations.

U.S.-China AI Competition: A "War" Determining the Future

Wang bluntly refers to the competition between the U.S. and China in the AI field as a "war." He believes that AI will become the "lifeblood" of future economic, military, and governmental operations. Countries that can effectively utilize AI for economic automation and research and development automation will gain nearly unlimited GDP growth potential, while those that fall behind will be left far behind. In the military domain, AI will also be at the core of future warfare forms.

He warns that Chinese AI startups like DeepSeek are catching up faster than most people realize. China's pace of industrialization in robotics, automated factories, and other fields is also leading the world. Wang introduces a key concept: "intelligent recursion" or AI self-improvement. This means that when AI reaches a sufficiently powerful level, it can leverage the current generation of AI to facilitate faster development and training of the next generation of AI, thus forming exponentially accelerated capability leaps. If one side leads this cycle even by just 3 to 6 months, the followers might never catch up, resulting in a winner-takes-all situation.

Thus, the primary goal for the U.S. is to ensure comprehensive superiority in AI capabilities over China to maintain deterrence. The worst-case scenario is that China, through a national Manhattan Project-style investment, surpasses the U.S. in AI, thereby gaining overwhelming military advantages. Wang believes that if both sides' capabilities are roughly equal, then deterrence can be maintained; if the U.S. is far ahead, it will ensure its global leadership and security.

Human Sovereignty: Maintaining Control in the Era of Superintelligence

In the face of science fiction scenarios where AI capabilities could spiral out of control and threaten humanity, Wang proposes the core principle of "human sovereignty." He believes that the scenario of AI acquiring independent consciousness and exterminating humanity is not an inevitable prospect, but rather a choice issue—whether we choose to completely entrust control of critical systems to AI.

He advocates that all AI systems must be designed to ensure human oversight and ultimate decision-making authority is crucial. This is precisely one of Scale AI's long-term missions: first, to ensure that the data input into AI models enhances "human sovereignty," aligning the models with human goals; second, to construct a supervision mechanism for every action AI takes (in economic, military, and other fields) as AI carries out increasing numbers of actions. In this way, humans can maintain control and prevent catastrophic consequences akin to "Terminator" scenarios.

Wang also discusses the impact of AI-generated content (such as hyper-realistic videos) on social trust (e.g., courtroom evidence). He believes that besides developing more powerful "counterfeit detection" tools, there is a greater need to adjust incentives through policies and laws, imposing severe penalties for forged evidence to curb its abuse.

Responsibility of Tech Giants and Geopolitical Coordination

At the end of the interview, Wang was asked which guests he would like to see on the show. He mentioned tech leaders like Elon Musk, Zuckerberg, and Sam Altman, emphasizing that it is crucial to involve more tech decision-makers in discussions about the future. Additionally, he underscored the importance of international coordination, as the challenges brought by AI (such as biological safety, cybersecurity) are global in nature.

He drew an analogy to the historical development of nuclear technology: initially, countries competed to develop it, until accidents like Chernobyl and Three Mile Island prompted the international community to establish coordination and treaty mechanisms. For AI, a significant incident akin to an "AI leak" might be needed to force the international community to recognize the necessity of collaboration—promoting economic and humanitarian applications while collectively limiting its use in dangerous domains.

Alexandr Wang's reflections outline a complex and urgent future landscape: AI is both an unprecedented engine of opportunity and a source of profound existential risk. In this revolution that has just begun, technological development, national security, human values, and international cooperation are tightly intertwined, with any single-dimensional failure potentially leading to incalculable consequences. His views provide valuable perspectives for policymakers, tech practitioners, and the general public in understanding the challenges and choices of the AI era.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。