Can someone explain how Cardano would address this problem described in the linked article below:
On-chain lending has become the most popular decentralized finance (DeFi) application today, with over $600M in loans originated this year across MakerDAO, Compound, and dYdX. On-chain lending has the potential to disrupt traditional secured lending. But it seems it may do more than that: it might also disrupt proof of stake consensus.
Proof of Stake (PoS) in an alternative to Proof of Work in which a blockchain is protected by staked cryptoassets instead of by hash power. Many of the major networks launched in the last year have been PoS networks (Tezos, Algorand, Cosmos, etc.), and many more are due to arrive in the next year.
A PoS system is secure when there are lots of coins actively staked for the network. In most PoS algorithms, so long as 2/3rds of all the staked assets are owned by honest actors, the blockchain will be secure.
Now imagine you are an attacker trying to break a PoS system. How would you go about it?
At a high level, there are two avenues of attack: you could accumulate 1/3rd of all of the outstanding stake, but that’s hard and expensive. The second approach is that you could convince the current set of stakers to stop staking and then take over the much cheaper network.
The second approach sounds attractive in principle, but how could you get the current set of stakers to stop staking? Here’s a simple way: offer them more attractive yield elsewhere.
PoS only works if stakers are incentivized to stake, and they’re only incentivized to stake if the rewards are big enough. But if they can get better returns elsewhere, then you should expect a rational staker to unstake their assets and put them wherever they earn a higher return. If this siphons demand away from the staking, the network becomes less secure.
In a very literal sense, on-chain lending markets directly compete with staking — meaning they directly compete with the protocol being secure!
You probably get the intuition that there’s an important interaction here we need to understand. But how exactly does one analyze something like this?