Written by: Techub News整理

Introduction

In the latest episode of the Network State Podcast, Ethereum co-founder Vitalik Buterin (V 神) engaged in a deep dialogue lasting 74 minutes with the host. This conversation took place just as Ethereum is about to undergo another important hard fork (the Pectra after the Dencun upgrade), amidst a critical juncture of flourishing Layer 2 ecosystems, maturing ZK technology, and the global wave of AI. Vitalik Buterin (V 神) is not only the soul of Ethereum but also one of the most influential thinkers in the fields of cryptocurrency and cutting-edge technology. In this interview, he systematically elaborated on the transformation of Ethereum from "concept" to "usable infrastructure" and shared his latest thoughts on grand topics such as AI, biosecurity, and startup societies, providing valuable perspectives for understanding the trend of the next stage of integration between cryptocurrency and technology.

Summary

• The path of Ethereum's scalability is clear: Layer 2 has achieved around 250 TPS, and after the Pectra upgrade, the number of Blobs will double. In the next year, it is expected to increase L2 TPS to thousands through technologies like PeerDAS. There is also a roadmap for L1 to achieve a tenfold increase in Gas limits by 2026 through a series of EIPs.
• DeFi security is significantly enhanced: Statistics show that the risk of randomly choosing a DeFi protocol suffering losses due to hacks has dropped to about 0.53%, with even lower risks for mature protocols. With improvements in wallet infrastructure like social recovery, multi-signature, and account abstraction, user experience and asset security are rapidly approaching traditional financial levels.
• AI carries both risks and opportunities: Vitalik Buterin (V 神) has expressed an increased concern about AI's "doomsday risk," believing that progress is faster than expected while global political cooperation is poorer than anticipated. However, he pointed out that AI currently resembles "augmented intelligence" rather than "autonomous intelligence," and the continued existence of prompting and the friction of physical world experiments are important limiting factors.
• Startup societies (like Zuzalu) are significant experimental fields: A two-month, approximately 200-person offline community acts as a real "full-stack incubator" for governance models, bio-defense technologies, and decentralized applications, transitioning from "talk" to "practice" to test the feasibility of new civilization concepts.

The Return of Ethereum: From Conceptual Totem to Mature Infrastructure

Vitalik Buterin (V 神) pointed out at the outset that Ethereum is at a critical moment: all the technological pieces long discussed are finally in place, making large-scale applications possible. He elaborated on this "return" from three core dimensions—scalability, security, and privacy.

Regarding scalability, historical lessons (such as 2017's CryptoKitties and 2021's DeFi summer) indicate that when demand surges against a fixed block capacity, exorbitant Gas fees kill the user experience. Today, Layer 2 solutions collectively handle around 250 TPS of transactions. The upcoming Pectra hard fork will double the number of Blobs, pushing L2 capacity to about 500 TPS. More importantly, technologies like PeerDAS (Data Availability Sampling) provide a clear path to increase L2 TPS to thousands within the next year. For L1 itself, the research community is working on "super optimization," maximizing performance under clear constraints to maintain decentralization and network resilience. A series of EIPs planned for implementation in 2026 are expected to increase L1's Gas limits tenfold. After that, ZK virtual machines (ZK-VM) will pave the way for higher-level scalability. "From a scalability perspective, we have already achieved a tenfold growth, the question is how to go further and improve interoperability between existing components," Vitalik Buterin (V 神) summarized.

Security is key to whether DeFi can become a mainstream wealth storage tool. Vitalik Buterin (V 神) candidly stated that three or four years ago, he couldn't confidently recommend DeFi as a savings and wealth appreciation tool to ordinary people, as the core issue at that time was not whether to achieve an annual yield of 6% or 4%, but to avoid a loss of "-100%" of the principal. However, the situation has greatly improved. According to data he cites, the ratio of total losses caused by DeFi hacks divided by the total locked value (TVL) is around 0.53%. This means that the risk of losing funds due to hacks in randomly chosen DeFi protocols is only about 0.5%. While this number is still concerning, it is important to recognize: first, this includes all risk protocols, and for mature protocols like Aave, the risk is much lower; second, this ratio has been declining over time. Additionally, risks from personal operational errors (like losing a wallet) are being mitigated through wallet technologies he has advocated over the past decade, such as social recovery, multi-signature wallets, and account abstraction (ERC-4337). He cited that even if mainstream smart contract wallet frontends like Safe are hacked, users relying on alternative UIs or carefully checking transaction signatures remain secure. The infrastructure is continuously being "hardened," and optional solutions are increasing. "From an average user's perspective, the ability to obtain decentralization and asset security is rapidly increasing," he compared the panic caused by the 2017 Parity wallet hack with the perfect security record of today’s smart contracts (like Safe core contracts), highlighting that every technology initially seems crazy and high risk, but is increasingly maturing and becoming “completely fine.” He even proposed a more radical view: in some ways, mature DeFi may be becoming safer than the current Western financial system. The risk of your money "disappearing" after a year in a traditional bank (due to bank bankruptcy, asset freezes, or political risk) may not be less than 0.5%.

Privacy is the third matured dimension. Vitalik Buterin (V 神) mentioned that Ethereum now has mature privacy solutions, such as Railway and launched Privacy Pools. Even Tornado Cash has regained legal clarity in the U.S. (although it never truly stopped operating). He called for the release of relevant developers. Beyond financial applications, he was also impressed by the decentralized social protocol Farcaster. Farcaster has not only survived for years with multiple clients and an active ecosystem but has also proven that decentralized social networks can transcend the "honeymoon phase," continuously attract users, and create value—not just because of its decentralized nature but because it offers a networked environment for rational communication, networking with interesting people, and engaging in interesting activities.

Vitalik Buterin (V 神) emphasized that Ethereum has experienced "rapid maturity improvements" in infrastructure and technology over the past few years, which can easily be underestimated because "improvements in maturity often feel like nothing has happened." However, it is this reliability of "no longer failing" that enables upper-level innovations. Just like building blocks, the underlying blocks (like Python, Django) must be 100% reliable and "boring" for developers to safely try a risk point on top. Today, many components of Ethereum and its ecosystem are reaching this reliable level of "okay, it works, I can take the next step."

Beyond Finance: Grand Narratives of AI, Bio-defense, and Startup Societies

After discussing Ethereum's technical advancements, the conversation shifted to the broader forefront of technology and society. Vitalik Buterin (V 神) shared his profound thoughts on artificial intelligence, biosecurity, and the experiments of "startup societies."

AI: Increased Risks, but Constraints Remain

When asked if he still considers himself an "AI doomsday skeptic," Vitalik Buterin (V 神) responded: his level of concern has actually increased. The reason lies in the fact that AI's progress is faster than expected while the state of global political cooperation is worse than anticipated. He expressed skepticism about the feasibility of many classic "doomsday" solutions (such as international treaties), stating that the prospects for achieving global security agreements in a world with escalating geopolitical conflicts and high tariff barriers are bleak.

However, he has his own judgments about the specific forms of risk. He believes that a true "killer AI" resembles drones or robots with autonomous capabilities, rather than current image generation tools or text summarizers. The alignment (Alignment) problem in AI essentially involves aligning AI with its controllers (which could be a tribe or nation), rather than with the safety of all humanity. Thus, AI alignment does not equate to AI safety. The critical threshold of risk hinges on whether AI has crossed the "approximately human-level autonomy and generality."

A core constraint of current AI is "prompting." Vitalik Buterin (V 神) pointed out that a significant debate over the past two and a half years centers on whether prompting will persist, and the fact is that there is no sign of it disappearing. This means that current AI resembles "higher-order programming" or "English programming," serving as an amplifier of human intelligence rather than truly autonomous agents. As long as prompting exists, AI's capabilities are largely constrained by the human intelligence levels of those using it.

Yet he also acknowledged that technological breakthroughs could change this situation. For instance, the "Chain of Thought" training approach used by DeepSeek has updated his perspective. This method allows the model to generate a lot of reasoning chains, filter out the correct answer for re-training, forming a cycle, and demonstrates for the first time a potential path to superhuman intelligence through training without apparent constraints. Additionally, embodiment is another critical bottleneck. Purely digital AI cannot directly affect the physical world, while manufacturing, deploying, and maintaining large numbers of robots involve significant physical world friction and business model challenges. He believes that those predicting "curing cancer and aging by 2029" seriously underestimate the complexity of experimentation and the physical world.

In summary, Vitalik Buterin (V 神) believes AI is both overestimated and underestimated. What is overestimated is its short-term autonomy and certain sci-fi scenarios; what is underestimated is its long-term fundamental impact, as well as its current usage in a "lazy" and "tasteless" manner (like generating dull tweets). He predicts that the areas where humans ultimately outperform AI may precisely be those tasks that are the most "ineffable" and the hardest to train.

Bio-defense and DeSci: Neglected Survival Issues

The conversation naturally transitioned from the risks of AI to another existential threat: bio-risk, especially pandemics. Vitalik Buterin (V 神) used COVID-19 as an example, pointing out a "split personality" in our society's response: it has caused the deaths of millions globally while being downplayed by many as "just a big flu." Our media mechanisms can only process one "current big event" at a time, leading to the epidemic being prematurely "resolved" in public discourse, despite the virus continuing to mutate, spread, and cause long-term health impacts.

He is concerned about the next, more severe pandemic. Solutions cannot rely on the currently politicized and discredited Western public health system. He proposed two paths: one is to turn to China, where there is a functional research and clinical system with relatively less political interference; the other is to depend on the internet and decentralized science (DeSci). DeSci means individuals can self-quantify (like air monitoring, personal genome sequencing), experiment on themselves, share data, and aggregate analysis with like-minded communities, thereby mastering their health and together promoting the adoption of bio-defense technologies (such as 222-nanometer UV lights, air filters, better ventilation).

Zuzalu and Startup Societies: The Operating System for Civilization Experimentation

This is precisely where the value of the concept of "startup societies" lies. Vitalik Buterin (V 神) used Zuzalu (a temporary community lasting two months with about 200 people), in which he is deeply involved, to elucidate the significance of such experiments. It occupies a unique position on the "number-duration" axis: long enough (two months) to become the participants' "life" rather than a "vacation," and large enough (around 150 people, reaching Dunbar's number) to require a multi-layered social structure, yet small enough to be practically organized and managed.

At this scale, communities can act as a "full-stack incubator" or "test user group," transforming new ideas about better governance and cutting-edge technologies (like bio-defense) from "online talk" into "real experiments." Unlike traditional "central-radiating" product models (e.g., Estonia's e-residency program), a true community is a net structure connected among members. This connectivity allows the community to generate internal demand, providing a valuable real testing environment for new products and services. Zuzalu has already spawned several derivative communities (like Zu Thailand, Zu Georgia), forming an ecosystem exploring new lifestyles and governance models.

Vitalik Buterin (V 神) drew an interesting analogy between cryptocurrency and the longevity biotechnology movement: both challenge the fundamental premises of traditional systems. Traditional finance considers slight inflation to be good, while traditional medicine assumes health decline is inevitable with age. In contrast, cryptocurrency and longevity communities start from first principles, questioning: can we avoid going to zero and instead move towards infinity?

Selected Q&A: Anonymity, ZK, Geopolitics, and Ethereum's North Star

During the audience Q&A session, Vitalik Buterin (V 神) provided direct responses to a series of sharp questions.

Regarding anonymity: When asked if he regrets not creating Ethereum anonymously like Satoshi Nakamoto, he stated that it was impossible at the time since he had already been active for two years as a writer for Bitcoin Magazine, leaving behind too many traces online. Moreover, Ethereum's early launch largely relied on the reputation and trust he had established. He agrees that pseudonymity itself is a form of decentralization (since real names are tied to government databases) but noted that there are various ways to achieve decentralization and that the vision of "network states" should be realized through nurturing many different startup societies rather than relying on individual pseudonymity.

Regarding zero-knowledge proofs (ZK): He believes that ZK technology is severely underestimated as it opens up a whole new design space, allowing users to establish and transfer verifiable reputation while maintaining privacy. This addresses the trust and spam problems of the early internet where "nobody knows you are a dog online." For early career choices, he suggested focusing on fields like ZK, biotechnology, and other areas that have not yet reached explosive moments.

Regarding geopolitics and dollar hegemony: In response to the question of "Most stablecoins are dollar-denominated, most Bitcoin is held by U.S. entities. Does crypto victory equal U.S. victory?" Vitalik Buterin (V 神) believes the future is full of uncertainties. Geopolitics, domestic politics, and AI are breaking the stable pattern of the past 80 years. Balaji Srinivasan (the host) proposed a more radical theory of the "DeFi Matrix": once assets are on-chain, they enter a global, low-friction exchange network at internet speed, similar to how local newspapers lost their geographic monopoly when moved online and had to compete ideologically on a global scale. Fiat currencies will face the same fate, with many weak fiat currencies likely to vanish as asset competition shifts from geography to characteristics (such as censorship resistance, programmability). The internal divisions within the U.S. mean that "U.S. victory" is not a given, and the internet, as a surviving infrastructure, has attributes that transcend its birthplace.

Regarding Ethereum's North Star: Vitalik Buterin (V 神) believes Ethereum's goals are dual: first, to be widely used, and second, to maintain its fundamental advantages over traditional systems (such as decentralization, security, censorship resistance). Sacrificing either side is unacceptable. Technically, through L1 optimizations (like delayed block execution, access lists), L2 scalability (more Blobs, PeerDAS), and ultimately using ZK/STARK proofs within L1, there is a clear path to achieve both scale and decentralization simultaneously. This requires rethinking some concepts from the Satoshi era, particularly utilizing the "asymmetry of creation and validation" so that a few nodes bear heavier computational loads to generate proofs, thus making massive validation extremely lightweight, achieving large-scale expansion without significantly raising entry barriers for nodes.

Regarding the innovation center: When asked if Silicon Valley is losing its status as the tech innovation center, Vitalik Buterin (V 神) believes it depends on the industry. For AI, Silicon Valley still holds value, but he suggests visiting for a few weeks every year to gain major benefits while avoiding "breathing in the madness." Cities like Berlin have strong, true-to-purpose crypto communities and top talent (like many Ethereum core developers and the Safe team). Combining the internet with offline hubs, the possibility of being at the forefront in non-global superstar cities today is much greater than it was five years ago.

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。