Preliminary Studies on decentralized communities in Conflicts and Societal Prosperity
University of Washington
Sergii GrybniakKey Ideas:
- Research Focus:
- Comparative analysis of centralized vs. decentralized communities
- How they manage conflicts triggered by external crises (Russo-Ukraine war)
- Social capital dynamics in Web3/NFT communities
- Airdrop hunters in decentralized communities
Study 1: GitHub (Centralized) vs. Aave (Decentralized DAO) During Russo-Ukraine Crisis
- Context (2022):
- War split developer communities into different camps
- Heated debates: Should we restrict access? Donate funds? Take political stances?
- Previous research only focused on internal conflicts (code disagreements, workflow disputes)
- Question: How do fundamentally different governance structures manage conflicts from massive external shocks?
- GitHub (Centralized):
- Staff has ultimate authority to modify content, lock discussions, make final decisions
- Clear hierarchical structure
- Timeline: Discussion about cutting Russia off → community responses → staff selected best answer → staff cancelled best answer → official statement → discussions locked
- Result: Centralized process triggered conflict transformation
- Discussions turned into affective conflict (personal attacks, identity-based arguments)
- Sentiment analysis showed increasing negative, subjective, personal sentiments
- 26 Polish, 42 Russian, 37 Ukrainian users felt compelled to reveal geographic locations
- Disclosure triggered identity-based attacks
- Members frustrated by bans without community consensus
- Aave (Decentralized DAO):
- No single authority can override community consensus
- Blockchain-based voting, token holders collectively determine outcomes
- Timeline: Proposal for donations from treasury → community responses → snapshot voting created → voting process → results announced → approval/withdrawal → closed in 10 days
- Result: Voting mechanism channeled disagreement into structural decision-making
- Rational, objective discourse persisted throughout
- Participants remained pseudonymous (blockchain addresses)
- Focus on argument merit rather than national/ethnic identity
- No transformation into interpersonal arguments
- Why Blockchain Makes DAOs Better at Conflict Management:
- Direct Financial Incentives:
- GitHub has reputation/badges (no direct incentive)
- Aave proposal passed → token price increased $36 in 2 days
- Immediate, tangible value (not just abstract reputation)
- Token holders benefit when good proposals pass
- Creates powerful incentive for thoughtful participation
- True Anonymity:
- GitHub users revealed locations, triggering identity-based attacks
- Aave participants remained pseudonymous (blockchain addresses only)
- Focus shifted to argument quality, not nationality
Study 2: Social Capital in NFT Communities (Nouns DAO Case Study)
- Research Question:
- How is social capital characterized in Web3 communities?
- Six dimensions: Network structures, collaborations, inclusion, communications, trust, empowerment
- Why Nouns DAO?
- Perfect extreme case for research
- Uses CC0 public domain license
- Founder quit after release, handed project entirely to community
- Pure decentralization
- Small part of derivative transactions automatically go to community treasury
- Benefits creators through blockchain smart contracts
- Methodology Challenge:
- Trust: How to quantify in contextual, subjective conversations?
- Inclusion: How to imagine global diversity from messy location data (unstructured addresses, emojis)?
- Previous studies used surveys/interviews (qualitative only)
- Solution: Delphi-based Human-LLM collaboration
- Human experts independently analyze data sample
- GPT does same work
- Facilitator collects outputs, normalizes, shares back to groups
- Feedback loop until consensus reached
- Calibrates LLM to human expert judgment
- Deploy LLM to analyze 2M tweets + 200K blockchain transactions
- Key Findings:1. More Decentralized Structure:2. Trust and Inclusion:3. Empowerment (Different Type of Success):
- Open bazaar vs. top-down cathedral
- Diverse, many-to-many communications (not one-to-many)
- Rich clubs (influential nodes) don't form exclusive connectivity
- Influential users engage broadly across network
- Healthier information flow
- General Twitter users discussing Nouns: Nearly 50% use cyber locations (protect privacy)
- Core Nouns community (using avatar/holding NFT): Over 50% share real-world regions
- Interpretation: Counterintuitive self-disclosure = strong signal of high trust and psychological safety
- Social connections have little/no correlation with direct trading gains
- Challenges idea that crypto communities are just "get rich quick"
- Thriving creative ecosystem supported by: a) CC0 License (Copyright Decentralization): b) Blockchain Smart Contract Royalties:
- Unleashed abundance of free derivative creations
- Extensive transaction networks for Nouns NFTs AND derived NFTs
- Part of royalties from derivatives flow to community treasury
- Treasury continues growing after founder left
- Members vote and fund new projects
- Creates self-sustaining ecosystem
- Shifts value from short-term speculation to long-term sustainable collaboration
- Theory of Human Activities Framework:
- Labor: Repetitive, driven by necessity/profit (most Web2 engagement)
- Work: CC0 + blockchain royalties empower creation and self-expression (diverse derivatives)
- Action: Owning governance = ability to participate in collective decision-making (shapes public life, missing in centralized systems)
Study 3: Airdrop Hunters (Sybil Attackers)
- Problem:
- Projects do airdrops to increase user base and decentralize
- People create fake identities to claim multiple airdrops (Sybil nature of blockchain)
- Harms ecosystem integrity
- Research Published:
- Kai 2003: Analyzed ParaSwap (DEX) airdrop
- Big data patterns of airdrop hunters
- Categorized into different groups
- Identified behaviors during and after airdrops
- Triple W 2024: Built GNN system to identify patterns
- Analyzed Blur airdrop Season 1
- NFT information + airdrop address transactions + post-airdrop activities
- Neural network to automatically identify patterns in future airdrops
- Current Focus (2024/2025):
- Reach consensus on airdrop hunter standards
- Define with numeric standards
- Accumulate Sybil addresses across systems (Ethereum, Solana, etc.)
- Purpose:
- Accelerate research (provide data for researchers)
- Industry reference for potential airdrop hunters
- Website with progress and database access available
- Challenges:
- KYC in Decentralized Systems:
- Centralized exchanges (Binance Alpha) can do KYC
- Decentralized communities have no centralized third party
- Anyone can generate address by creating key
- Many communities use stake-based systems (money = real person proxy)
- Dual Incentives:
- Projects don't like airdrop hunters (don't want tokens going to Sybil attackers)
- BUT projects still want high user numbers to prove traction
- Use inflated numbers to convince centralized exchanges (Binance) to list tokens
- Three-party battle: Airdrop hunters pretend to be good users, projects need user numbers, centralized exchanges need to identify real users
- Solution: Develop skills, definitions, databases as reference for all parties
- Academic Position Available:
- Lead researcher: Dr. Hongzhe Chen (CKB Eco-Fund)
- Open to work in academia
- PhD student openings at University of Washington in decentralized computing
In the context of post-conflict digital reconstruction, decentralized communities offer innovative pathways to enhance resilience, trust, and societal prosperity. This talk introduces our empirical studies on decentralized autonomous organizations (DAOs) and NFT ecosystems. First, we compare Aave (a DAO) and GitHub (a centralized autonomous organization, CAO)—both developer-driven communities—in managing conflicts triggered by the 2022 Russo-Ukrainian Crisis. Using co-occurrence networks, sentiment analysis, and user activity data, we reveal that DAOs mitigate conflicts more effectively through blockchain-enabled voting and cryptocurrency incentives, reducing affective tensions and fostering consensus. Second, we examine social capital dynamics in the decentralized Mfers NFT community, analyzing data from tweets, users, transactions, and addresses. Employing network analysis, NLP, and a novel Delphi-based human-LLM collaboration, we map six dimensions: decentralized structures, inclusive collaboration, diverse communication, high trust, and empowerment via royalties and derivatives. While promoting innovation and equity, challenges like wealth inequality and ethical issues emerge, highlighting DLTs' potential for holistic societal prosperity. Integrating these insights, we advocate a DLT-driven architecture to support autonomous ecosystems.
Bio: Dr. Wei Cai is a tenure-track Assistant Professor of Computer Science and Systems at the School of Engineering and Technology, University of Washington, Tacoma, WA, USA. He is now leading the Decentralized Computing Laboratory. He holds a Ph.D. in Electrical and Computer Engineering from The University of British Columbia (2016), an M.Sc. in Electrical Engineering and Computer Science from Seoul National University (2011), and a B.Eng. in Software Engineering from Xiamen University (2008). Prior to joining UW, Dr. Cai was an Assistant Professor of Electrical and Computer Engineering at The Chinese University of Hong Kong, Shenzhen, China. He has also conducted research visits at Academia Sinica (Taiwan), The Hong Kong Polytechnic University, and the National Institute of Informatics (Japan). Dr. Cai has co-authored over 100 peer-reviewed journal and conference papers and has received 7 Best Paper Awards. His research focuses on decentralized computing, with emphasis on mechanism design, social computing, multimedia, and applications. He serves as an Associate Editor for IEEE Transactions on Mobile Computing (TMC), ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM) and IEEE Transactions on Computational Social Systems (TCSS), and previously for IEEE Transactions on Cloud Computing (TCC). Dr. Cai is a Steering Committee member for ACM NOSSDAV, where he served as TPC co-chair in 2023, Associate Chair for ACM CHI'26, and has been an Area Chair for ACM MM since 2023. He is a Senior Member of IEEE and a member of ACM.