Dialogue between Vitalik and Suji: Why have decentralized social products failed?

Editor | Wu Says Blockchain

This issue features a Chinese AMA jointly held by Wu Says and Mask Network on January 23, with participants including Ethereum co-founder Vitalik Buterin, Mask founder Suji Yan, Wu Says editor Colin Wu, and other community representatives. The discussion primarily revolved around the chain reaction triggered by Twitter's concentrated ban on "mouth-rolling platforms," delving into the real dilemmas and new opportunities of decentralized social media.

Vitalik analyzed the fundamental reasons why most decentralized social products have failed, from perspectives such as the difficulty of breaking through network effects, misaligned incentive mechanisms, and excessive financialization. He emphasized that products should be redesigned from the perspective of "the problems of social itself" rather than "the overlay of crypto finance." Suji pointed out through case studies of products like Lens and Farcaster that user migration is a slow, phased process that requires the coexistence and collaborative advancement of centralized and decentralized forms. The discussion also extended to the potential integration of social protocols with wallets, prediction markets, AI, and community governance, suggesting that if the goal is "to improve the quality of discussions and approach facts and consensus," there remains long-term exploratory value in decentralized social media.

The personal views of the guests do not represent the views of Wu Says and do not constitute any investment advice. Please strictly adhere to local laws and regulations.

Audio transcription was completed by GPT and may contain errors. Please listen to the complete podcast on platforms like Xiaoyuzhou and YT.

Vitalik's Renewed Focus on the Realities of Decentralized Social Media and Two Major Reasons

Colin: Recently, Twitter's concentrated ban on "mouth-rolling platforms" has reignited discussions about decentralized social media. Vitalik, what are your thoughts on this event and the overall trend?

Vitalik: I have been paying attention to decentralized social media for many years, and I have used Farcaster quite a bit from 2023 to 2025. Recently, I have been more clearly focused on this direction for two main reasons.

First, the problems with Twitter are becoming increasingly apparent. Whether in crypto, politics, or public issues, the overall quality of discussions continues to decline, and it is no longer suitable as a center for global public discourse. However, the biggest challenge for new social platforms remains the network effect; no matter how good a product is, it cannot succeed without users.

Currently, there are two relatively viable paths in decentralized social media: one is to split the protocol layer, placing accounts and content on a shared protocol, allowing different teams to create clients so that new users can see the complete network as soon as they join; the other is an aggregation model like Firefly, which places Twitter, Farcaster, Lens, etc., in the same interface, gradually guiding users to migrate without leaving Twitter. Both approaches essentially aim to alleviate the problem of network effects.

The second reason is that the technology has matured. Over the past three years, whether it's Firefly, Farcaster, or the broader crypto infrastructure, privacy technologies, and communication tools, the user experience has significantly improved.

With these conditions combined, along with a series of recent events such as the Kaito ban and changes in Farcaster and Lens, I believe now is a time to push for decentralized social media again and worth a serious revisit.

Suji Reflects on Early Cooperation with Twitter and Differences in Centralized Platform Governance

Colin: Suji, what are your thoughts on the recent event of Twitter's concentrated ban on "mouth-rolling platforms"?

Suji: Our cooperation with Twitter dates back to 2018, during Jack Dorsey's tenure, when Twitter was very open to crypto experiments, including crypto tweets, on-chain transactions, and later turning the first tweet into an NFT for auction. These reflected an early governance style of Twitter that was more progressive and rule-oriented.

The changes became very apparent after Elon took over. Although he also emphasizes freedom of speech, his approach is more about strong enforcement, and this governance difference is directly reflected in product and banning strategies. Based on this judgment, I was not surprised to see products like Kaito being banned during the Elon era.

Projects like Lens and Farcaster initially had a strategy that somewhat involved "confronting Twitter head-on." Personally, I prefer Jack's approach from back then: don't rush to replace, but patiently and gradually migrate user behavior and traffic out.

I also noticed that exchanges have been trying to create social products in recent years, such as Binance Square and OKX Orbit, but they are very cautious about deeply integrating with Twitter, which also indicates the uncontrollability of centralized platforms.

Therefore, I am more optimistic about decentralized social media as an integration layer, but this will definitely be a long-term process. A more realistic path is to transition from centralized platforms to a "semi-decentralized" social form, and then gradually move towards complete decentralization. This path is very similar to the evolution of DeFi and prediction markets. Currently, multiple events are happening in a short time, which is essentially a phased release after long-term accumulation.

Core Issues Behind the Long-term Difficulty of Decentralized Social Media Success

Colin: There is a strong demand for decentralized social media, but in recent years, whether it's Friend.tech or the recent Kaito and Farcaster, it's been difficult to achieve PMF. Why has this direction been so hard to succeed?

Vitalik: I think there are two main reasons.

The first is that the network effect is too strong. In the past decade, almost no new social platform has truly grown large; Twitter remains one of the core platforms, indicating that the barriers to entry in the social space are extremely high.

The second reason is that many projects have misunderstood the problem, treating "crypto finance" as the solution to social issues, jumping straight into SocialFi, adding tokens, and adding trading layers. But starting from social itself, the truly critical issue is often creator incentives: how to ensure that newcomers and high-quality content receive continuous rewards.

Such mechanisms have been tried in crypto for over a decade (Steemit, Bihu, various Creator Tokens), and many failures are because the incentives actually reward existing influence and speculative demand rather than content quality. In contrast to Substack, which allows high-quality creators to earn income over the long term, many tokenized platforms often only reward big V influencers, exchange founders, and others with "existing social capital."

Therefore, to create a good decentralized social media, the starting point should be to clarify which social problems need to be solved, rather than thinking first about how to overlay financial layers.

Experiences and Regrets of Early Social Products in the Chinese Community: Migration, Governance, and Patience

Colin: The Chinese community has actually tried to combine social media and crypto for a long time, such as Bihu and Mirror. Bihu once gathered a large number of creators but ultimately shut down; Mirror temporarily replaced public accounts and Medium, with many institutions and authors writing long-term but halted due to team issues. These indicate that the demand has always existed, but a true PMF has never been achieved. Suji, what are your thoughts on this issue?

Suji: If we only look at the Chinese-speaking community, migration has actually been happening, albeit very slowly. In the early days, everyone mainly used Weibo and WeChat. Despite the existence of real-name systems and security risks, due to habit, very few people left. It wasn't until a batch of crypto accounts and media accounts were concentratedly cleaned up that everyone migrated to Twitter on a large scale, while WeChat groups also shifted to Telegram.

Twitter itself is not decentralized, but mechanisms like Community Notes allow the community to correct the mistakes of authoritative figures, reflecting a certain spirit of decentralized governance, which is very appealing to many. However, later on, people also found that Twitter still misfires on creators and lacks transparency in communication, leading some to try alternatives like Firefly, rather than immediately leaving Twitter completely.

Historically, the transitions from Weibo to Twitter and from WeChat to Telegram have been driven by regulation and product experience. Today, decentralized social media has yet to reach a similar critical point and can only continue to be pushed patiently.

Additionally, neutrality is crucial. For example, the early growth of Farcaster largely benefited from Coinbase's support, but if it is perceived as serving only one company or a small circle, it may deter other participants. A truly sustainable social infrastructure must engage participants from different factions.

I believe now is a noteworthy time, as a series of events are happening intensively, indicating that migration has already begun, but to truly complete it may still take a few years.

Lens's Multi-Platform Cooperation Approach and Social Opportunities in Prediction Markets

Colin: Suji, what is the current situation of your cooperation with Lens?

Suji: We have been shareholders of Lens for a long time. Over time, I have become increasingly certain of one thing: whether it's Lens or Farcaster, there is no need to go it alone or only embrace one ecosystem. The crypto social market is still very small and must be sufficiently open—cooperating with all wallets and all chains, and even to some extent being able to access Twitter's data and social relationships, is necessary for growth.

We talked with Stanley for a long time and ultimately decided to formally step in and become the new management of Lens. I don't want to describe it as an "acquisition"; it feels more like a cooperative alliance, not a case of one party swallowing another. That period also coincided with a series of events like Kaito and Farcaster, but those were just coincidences. I actually hope everyone doesn't focus on competition; the market is too small, and collaboration is what matters most.

Colin: So what will Lens do next?

Suji: The first step is that I hope all wallets can post directly using Lens, regardless of which chain or exchange the wallet is from. The second step is that I believe "prediction markets + social" is severely underestimated. Places like Polymarket see many betting behaviors that inherently spark discussions, jokes, and interactions, but the current comment section experience is very rough, and almost no one has turned it into a true social product, which is an opportunity.

Broadly speaking, it's not just prediction markets; trading behaviors, comment sections, content incentive mechanisms, and even the "social expression" of Memecoins can be interconnected—for example, sending a coin with the message "lunch money," which is essentially like a WeChat red envelope. Our approach is to take it slow, starting from wallets, then moving to prediction markets and trading comment sections, gradually connecting these real social behaviors over the long term.

The Possibility of AI + Community Annotations Enhancing Discussion Quality

Colin: Besides decentralized social media itself, Vitalik, have you noticed any new and interesting trends this year? For example, could combining with prediction markets be a direction?

Vitalik: I think this is a very promising direction. This year on Twitter, a relatively positive change has been the introduction of Grok. Now, when someone posts obviously inaccurate or even extreme views, people can directly ask Grok, which often points out what is wrong and what is right, thereby improving the quality of discussions to some extent.

Of course, Grok is not perfect. Twitter now also has Community Notes, but the biggest problem is that feedback is too slow, and annotations often take a long time to appear. I have been thinking about how to make "correction" happen faster, and currently, it seems that prediction markets could be a feasible tool.

For instance, when someone makes an extreme judgment, a prediction market could be initiated directly, and through the probabilities given by the market, one can quickly see how unlikely that statement is to occur. In many cases, this can effectively alleviate emotional and irrational discussions.

However, some prediction markets are currently deviating from their original intent, placing more emphasis on sports betting or gimmicks rather than "helping to approach the truth." Therefore, I believe that if we can combine AI, mechanisms similar to Community Notes, prediction markets, and search into one interface, with a clear goal of "improving discussion quality," this would be a very worthwhile direction to explore.

Additionally, the "secondary prediction market" proposed by Glen Weyl, which combines prediction markets with secondary voting, identity, and governance, is also very suitable for experimentation on social platforms. Overall, as long as the goal is clear, focusing on "helping people better understand reality and the future," the combination of AI, prediction markets, and social media has long-term value.

The Rationality of Coexisting Prediction Markets in Different Cultural Contexts

Forrest: In prediction markets, people's opinions and judgments are an important part of price discovery, not just funding and liquidity at play. In the future, we may not need just one prediction market. The West already has Polymarket; does the Chinese context also need a prediction market that represents local culture and user structure? Even for the same event, prices and judgments formed under different cultural backgrounds may differ, and this difference itself is valuable and worth discussing.

Vitalik: This is an open question, and the key lies in whether there are essential differences in the needs of different cultural communities for prediction markets. There is currently no definitive answer. However, I am very certain that this field needs more competition and experimentation. Whether it's different market forms, different application scenarios, or combining prediction markets with prediction-based governance (Futarchy), all are worth exploring in parallel. Ultimately, which model will prevail is unknown, and precisely because of this, more directional attempts are needed.

Interoperability Issues in the Combination of Social Protocols and Wallets

0xLuo: The combination of social protocols and wallets is increasing, whether in decentralized social media or centralized platforms like Twitter, which are adding trading and financial functions. With Firefly also integrating wallet and prediction market functions, what are your thoughts on the trend of "social products becoming wallet-like and wallet products becoming social"?

Vitalik: The key premise is that the underlying protocol of social media should remain as simple and neutral as possible. If the underlying layer is overly customized and filled with too many functions, it will weaken the network effect and hinder collaboration between different products. I do not wish to see every wallet develop a social network that is incompatible with others.

One important success experience of Ethereum is "wallet interchangeability": users can freely change wallets and migrate private keys. This openness is equally important when combining social and wallet functions; interoperability must be prioritized.

The functions of wallets will certainly continue to expand, especially in the context of increasing importance of privacy, where wallets not only protect assets but also data. At the interface level, integrating social and financial functions together is reasonable, such as tipping, simple decision-making, ENS, etc. However, it should be noted that many social users are not native on-chain users.

Therefore, a more realistic design is to first provide built-in wallets to lower the entry barrier while ensuring that users can migrate their accounts and assets at any time. This "low barrier to entry + freedom to migrate" model is key to whether the combination of social and wallets can go far.

Further Exploration of the Integration of Social Media, Prediction Markets, and DAO Governance

Instant Pig: I have been paying attention to the combination of Community Notes, social graphs, social identity, and prediction markets, and I am also working on prediction market-related oracle tasks. I would like to ask a question: will there be a scenario in the future where users can directly link content to a prediction market while scrolling through their information feeds, allowing them to participate on the spot? If Firefly does this, what advantages would it have over Twitter?

Vitalik: This deeper integration is very interesting and has many forms that can be tried. For example, under a post, instead of just liking or retweeting, there could be a direct "I agree / I disagree" prediction market, allowing everyone to see a probability given by the prediction market simultaneously. Experiments focused on the discussion itself are very worthwhile to continue exploring.

Instant Pig: What if this mechanism is extended to DAO governance, such as initiating predictions or bets within a small scope based on identity to assist decision-making? Would this type of prediction-based governance (Futarchy) be more suitable for implementation on decentralized platforms?

Vitalik: Currently, there is not a truly designed social platform for DAO or protocol governance; many governance discussions occur on Twitter or in private groups, which is problematic in itself. If the social interface is designed from the outset with the goal of "serving governance discussions," combining discussions, predictions, and voting as on-chain decision-making behaviors, this would be a very worthwhile direction for decentralized social media to explore in depth.

Neutral Evaluation of Centralized Companies Building Their Own Social Platforms

Colin: What are your thoughts on products like Binance Square? Some even joke that it might be one of the most active social products in the crypto space right now, although it is certainly not decentralized. I feel the content mainly revolves around trading coins, and the users are more "grassroots." Many might not use Twitter or Telegram but communicate directly within Binance. However, there are also issues of market manipulation, emotional manipulation, and even fraudulent content. What do you think about exchanges doing social media?

Vitalik: I have indeed looked at Binance Square a bit and know of its existence, but I haven't used it in depth, so it's hard to evaluate the content quality specifically. In principle, I don't think there's anything wrong with companies creating their own social platforms. In fact, having all people concentrated on one social platform is itself an unhealthy state.

If we can transition from "one centralized social platform" to "many centralized social platforms," and each platform better meets the needs and usage scenarios of its respective community, that would already be an improvement. Taking it a step further, if these platforms can connect with each other to some extent, for example, based on a common protocol that allows users from different platforms to still see each other's content and find more suitable communities, that would be even more ideal.

I am also thinking that if a high-quality, decentralized, and sufficiently neutral social platform really emerges in the future, even my own blog's comment section might directly use such a system. In other words, if decentralized social media can succeed, it is likely not to be "a unified product," but rather a large number of companies, communities, and individuals creating completely different interfaces and products based on the same underlying foundation for entirely different goals.

Suji: So do you hope that social media created by exchanges like Binance, OKX, Coinbase, and Bitfinex can eventually interconnect like Ethereum transfers? So that each doesn't have to create a completely incompatible system.

Vitalik: This question actually requires more detailed consideration. I do not believe that social media must achieve completely "zero-friction" interoperability. Sometimes, if a community wants to maintain its sense of boundary and set certain thresholds, that can be meaningful.

At the same time, interoperability itself remains very important. The key is not "whether to interconnect," but "at what level to interconnect and to what extent." This may require finding a more suitable balance between openness and community differences.

How to Attract Talent Back to Build the Blockchain Industry Again

Jocy: We have been discussing recently why applications like social media and gaming in blockchain have not truly succeeded. After entering the AI cycle, the scale of internet applications has rapidly expanded, and user attention has become more dispersed, leading many to start questioning whether decentralized social media is a false proposition.

In the current environment, our judgment is that directly creating an application that competes head-on with mainstream internet products has a very low probability of success. A more realistic path is to first create niche products aimed at crypto users, successfully operating within small circles, and then consider expansion, like the model of Polymarket. However, it is still more finance-oriented, and social and gaming products aimed at general users are currently very difficult.

Another feasible path is ToB, creating products that have real demand, PMF, and can generate revenue. In the AI field, many ToB companies are actually more likely to succeed.

I would like to ask Vitalik, in such a market environment, what kind of founders are more likely to create decentralized applications? And how can we encourage more excellent builders and talent to continue participating in this industry?

Vitalik: I think the reason blockchain social media and chain games repeatedly fail has a common cause: too many projects start from "we are crypto, we can add finance," rather than from the application itself.

The real questions to ask are three: What application are we creating? What is this application truly lacking right now? Can we solve these problems with existing technology? Many projects that seemed very successful in 2021 rapidly lost users in the bear market, fundamentally because users did not feel the product was good; they were just in it for the money. The same logic applies to social products.

Therefore, the next generation of founders who are more likely to succeed should be those who truly understand social media and user needs, even those who "understand social media but not blockchain" may be more suitable than those who "only understand blockchain." The same goes for blockchain games; understanding games is more important than understanding chains.

As blockchain gradually becomes infrastructure, it should play a more significant role in the background rather than being the product itself. Decentralization is not "on-chain + issuing tokens + speculation," but using blockchain as a reliable foundation for data and collaboration. If we start from specific use cases rather than from technology or tokens, we are more likely to create products that people are willing to use long-term.

Cultural Differences Between China and the U.S. in Crypto and Advice for Chinese Developers

Jocy: I would like to ask how you view the cultural differences between China and the U.S. in crypto. The U.S. has more of a protocol and developer culture, while China leans more towards exchange culture, with relatively few large protocols and infrastructures. After entering the AI cycle, the speed at which Chinese talent is leaving crypto has become noticeably faster. What advice do you have for Chinese developers? How can we encourage more people to continue participating in decentralized construction?

Vitalik: I have always believed that the greatest advantage of Chinese developers lies in product and user experience, especially front-end capabilities, which excel at providing high-quality experiences for large-scale users. This is particularly prominent globally.

Compared to ten years ago, the underlying infrastructure is now very mature, and there is no longer a need for every team to build L1 or underlying systems from scratch. Coupled with AI lowering the development threshold, I do not recommend that Chinese developers pursue complex, heavy-asset underlying protocols.

A more realistic path is to treat blockchain as infrastructure in the background, focusing on effectively utilizing existing protocols and innovating at the application and experience levels. You don't necessarily have to create new protocols; you can also create excellent clients or applications. Ultimately, what determines the success or failure of a product is still user experience and implementation capability, which is precisely where Chinese developers have the most advantage.