Reproduce “Bittensor Protocol: The Bitcoin in Decentralized Artificial Intelligence? A Critical and Empirical Analysis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Elizabeth Lui, Jiahao Sun
Researchers at FLock.io conducted an extensive empirical analysis of the Bittensor protocol, revealing severe centralization of stake and rewards, and weak correlation between miner performance and earnings. The study proposes and validates protocol-level interventions, such as an 88th percentile stake cap, which substantially enhance security against 51% attacks and improve incentive alignment.
Key insights
- Bittensor exhibits extreme centralization of stake, with the top 1% of wallets controlling a median of 90% of the total stake across subnets.
- There is a very strong correlation between staked TAO and rewards, but a very weak correlation between miner performance (trust score) and rewards, indicating misalignment of incentives.
- Over half of Bittensor subnets are highly vulnerable to 51% attacks, often requiring fewer than 1% of wallets to control majority stake.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.