Proof of Stake
The failures of a centralized "consensus" mechanism
Proof-of-stake is no alternative to proof-of-work. When deployed, proof-of-stake will run into failures in the long run, because it has fundamentally flaws that can not be addressed without proof-of-work.
The fundamental flaws are:
- No trustless mechanism for decentralized timestamping
- No objective truth
- No link to reality
- No true randomness
- No truly fair validator selection
- No validator turnover
- No way to bootstrap & fairly distribute coins
- No real-world external cost
- No game-theoretic stability in an adversarial environment
- No way to reduce imbalance over time; rich get richer
- No aligned incentives
Read the following to understand why proof-of-stake has and will always fail:
- What’s Wrong with Proof of Stake? by Bob McElrath
- Proof-of-Stake Fallacy by Eric Voskuil
- Proof-of-Stake & the Wrong Engineering Mindset by Hugo Nguyen
- Work is Timeless, Stake is Not by Hugo Nguyen
- Nothing is Cheaper than Proof of Work by Paul Sztorc
- Proof of Stake is Still Pointless by Paul Sztorc
If you are more on the academic side of things, here are some papers:
- Proof-of-Stake Is a Defective Mechanism, V. Sus (2022)
- On Stake and Consensus, A. Poelstra (2015)
- Distributed Consensus from Proof of Stake is Impossible, A. Poelstra (2014)
Some recent failures include:
- 2022-05-27 Solana’s blockchain clock loses track of time, now running 30 minutes behind (see Lesson 17 and Bitcoin is Time)
- 2022-06-19 - Solana: 1 vote, making up 90% of total votes, decided the fate of ~$270m in users assets
- 2020-03-18 - STEEM: adversarial takeover by Justin Sun, resulting in a hard fork that is now Hive (timeline, details).
For some not-so-recent failures, see this thread by Alex B.
Visit endthefud.org/PoS for a collection of articles that explain why proof-of-stake is a non-solution to the problems that Bitcoin solves. Also make sure to browse the rekt leaderboard to get an idea of the amount of value destroyed and stolen by those who peddle bad ideas.
Consider sharing it, translating it, or remixing it in another way.