Kyle is the #1 VC investor but on this thesis -- "FHE is best for Privacy" -- you're wrong:
Two problems with FHE: you need the *nodes of the chain* to do the heavy lifting of computing the FHE. Because FHE gives privacy, but not *integrity*. If you need integrity, you either need to trust the FHE operator (centralized setting, not blockchain), or *add ZK* to the mix, because ZK does prove integrity.
Second problem is the silicon and scale needed are pretty large, especially at scale (1000TPS). And you'd need to nodes to run it, hurting decentralization and scale.
ZK allows to offload massive computation to one, non-trusted, computer, and still process those transactions and put them onchain, no trust assumptions.
I love Zama, and love FHE. As co-inventor of a lot of the ZK tech deployed today (Zcash and StarkWare co-founder) I have immense respect for teams taking moon-math and making it a reality, like Zama.
But FHE ain't the silver-bullet you describe.
6/
b) is the most subtle and hardest to understand. This is where ZK fails
To understand why, let's consider the simplest privacy app: zcash (no defi).
When you submit a shielded zcash transaction, you produce a proof that says, roughly "I am sending coins such that my balance remains >0 after this transaction."
Well, if you aggregate 1000 of those transactions and then look at the state of the chain as an outsider, what do you know about the state? Nothing
Now imagine trying to do DeFi on top of that. How do you do DeFi if your transaction literally cannot see or interact with anyone else's assets
A bunch of teams have been trying to address this over the last decade, including Aztec, Aleo, and probably quite a few more that I cannot recall atm.
The fundamental challenge each of these teams has to contend with is the challenge described above. Basically, how do you design ZKPs to allow for selective information to be seen by the outside world (eg how much collateral is backing a loan)
Now consider being a DeFi developer. You have to not only design your DeFi protocol, you have to do so 1) in a partially hamstrung way, and 2) you have to understand how ZKPs work. Who wants to be the developer building a DeFi system that has 9 or 10-figures in it with all of these additional risks. This is scary stuff
Many of the zk DeFi teams have been working to try to make this stuff more understandable, but the underlying reality is just extremely difficult to contend with.
Moreover, this requires rebuilding every DeFi primitive from the ground up.
The fundamental challenge here is that DeFi as we currently know it *requires the ability to reason about globally shared state.*
Perhaps there is a way to rebuild DeFi from the ground up with selective reasoning, but I am extremely skeptical of that claim. And demonstrating that claim to the world at large in a way that everyone else will believe it will be a decade long endeavor given how much technical risk is there with dozens of bespoke zk circuits.
So what is FHE? FHE allows you compute over encrypted data. This has been considered the holy grail of cryptography for decades.
Thinking about private DeFi using FHE as the primary cryptographic construction is actually quite simple. You think about it the same as if it was transparent! It's just that magically, everything is not transparent, but you can still compute over it anyways
Yes, this is magic
7.16萬
253
本頁面內容由第三方提供。除非另有說明,OKX 不是所引用文章的作者,也不對此類材料主張任何版權。該內容僅供參考,並不代表 OKX 觀點,不作為任何形式的認可,也不應被視為投資建議或購買或出售數字資產的招攬。在使用生成式人工智能提供摘要或其他信息的情況下,此類人工智能生成的內容可能不準確或不一致。請閱讀鏈接文章,瞭解更多詳情和信息。OKX 不對第三方網站上的內容負責。包含穩定幣、NFTs 等在內的數字資產涉及較高程度的風險,其價值可能會產生較大波動。請根據自身財務狀況,仔細考慮交易或持有數字資產是否適合您。


