TL;DR
Mind's Restaking accommodates almost everything, including BTC, ETH, and even LST/LRT assets of all other network tokens
Mind business logic is similar to EigenLayer's AVS, but can also cooperate with Eigen and even all other networks
Mind is still in its early stages, so not many people can participate, but you can try it out, maybe there will be surprises in the future
Mind’s current main solutions focus on AI and Depin, +FHE, +Restaking narrative
Mind is like an abstract player. It is on the FHE track, but does not compete with other FHE projects. It is on the AVS track, but does not compete with AVS projects. It is on the Restaking track, but does not compete with Restaking projects. On the AI and Depin tracks, okay, serve them!
Participation experience link: https://dapp.mindnetwork.xyz/
What is Mind Network?
Mind Network is defined as a FHE Restaking Layer serving AI and PoS networks.
AI and PoS: This is what is served
FHE: technical security layer, which is the technical support, homomorphic encryption, and ensures fair verification
Restaking: Economic security layer, economic support, source of consensus
Understand it in the way of AVS, the left hand accesses the Restaking assets, and the right hand provides security consensus for AI and other networks
Main products:
Subnet: Subnet, a specific use case based on FHE
Remote Restaking: Remote staking, assets can participate in Mind restaking in the original chain
Mind Chain: Rollup based on Altlayer, mainly responsible for connecting Restaking and Subnet
FHE Bridge: Based on stealth address and CCIP, FHE cross-chain bridge is mainly used for B-side
Mind Lake: FHE privacy database, mainly serving the B-side
Investment and Background:
Completed $2.5 million seed round in 2023
Binance and Chainlink Ecosystem Incubation Project
Ethereum Foundation Grant
Testnet data (ended): 650,000+ active users; 3.2 million+ transactions
---
Mind Network Architecture
The whole can be divided into three parts
Economic input (Restaking layer): Receive LST/LRT assets, support multiple restaking tokens such as Ethereum and BTC
Intermediate architecture (Mind's own security layer + consensus layer): Based on FHE, encrypt the votes, encrypt the process, and build an overall FHE verification network
External output (subnet): Provides shared security based on Restaking and FHE for Depin, AI, PoS and other networks
Subnet
Mind Network’s solution is called Subnet. According to the introduction, Subnet is a verification demand network based on FHE and Restaking, which is similar to the upgraded version of AVS.
Each Subnet can customize its own tasks and logic, such as validator node requirements, reward rules, FHE-related functions, encryption and decryption, etc.
According to the current public information, the first batch of subnets should be mainly AI and Depin, such as Io.NET, Myshell, etc.
Remote Staking First of all, it should be made clear that Mind Network supports LST/LRT assets, but not original assets, such as BTC/ETH. The current version mainly supports Ether.FI, Renzo, Lido, and Stakestone's LSTRemote Staking. Remote staking provides a way to participate in staking with the lowest security assumptions, without the need for cross-chain, for example:
On the manta chain, use ETH to participate in Stakestone staking and get Stone
On Mind, switch directly to the Manta chain and click stake to stake Stone.
There is no need to cross-chain Stone to the main network.
The main benefits of participating in remote staking: assets are on the original chain, operations are simple, and security assumptions are low.
The current question is how the final rewards will be distributed and to which chain, which is also a question that needs to be paid attention to in the future.
AI and Decentralized AI
The most easily perceived AI feature by users is the model, which is generated by two key resources: computing power and data.
One is raw material, the other is power
At the computing power level, enterprises monopolize and it is difficult for small and medium-sized AI to seize computing power. Decentralization can utilize the idle computing power in the hands of users through token incentives, such as Depin
At the data level, data cleaning and labeling are upstream, and then they are provided to the model for processing and optimization. During the whole process, data faces multiple threats, such as storage, data calculation, data output, etc.
Regarding security issues involved in the data process, such as when talking with Chatgpt, we essentially establish a trust assumption and trust that OpenAI will not peek into our data.
Thus, decentralized AI was born, such as the cloud computing market, cloud training market, AI agent market, GPU computing market, prediction market, generative dialogue, etc. However, Crypto AI also faces some difficult problems.
Cloud computing power or Depin projects encounter fairness issues, and fairness is essentially a consensus issue.
To whom should the task be assigned, and how should the data of one’s own equipment be leaked and the contribution be fairly judged?
Problems facing decentralized data marketplaces like Bittensor
Miners train models, and validators vote to give scores. How can I be sure that validators are not colluding? How can I be sure that the model I am given is the best?
The data privacy involved in the whole process can be roughly divided into several parts:
Will user device data be leaked to the platform?
Will the data that needs to be calculated be leaked to users?
How can the platform prove its innocence and safety? ...etc.
So the great cryptography came: data encryption, FHE operation, ZK verification, MPC management authority
The ultimate utopia: end-to-end encryption of AI.
That is, AI doesn’t know what we are asking, nor does it know what it is calculating, but AI can output the answers we want.
Problems that FHE solves
The core of the PoS network is voting. The original intention of voting is to hope that the validator can independently verify and reach a consensus on the public network based on the results.
But in reality, there are very few networks that can be used for a large number of nodes like Ethereum.
Only when there are more validator nodes can we ensure that a small amount of cheating does not affect the overall situation, that is, solve the BFT Byzantine problem.
If there are too few nodes, cheating and manipulation will occur, such as rewarding follow-up investments and bribery.
Based on the problems of PoS network, Mind provides FHE-based security verification. Simply put, it encrypts the voting process to avoid the situation where the number of nodes is insufficient (not decentralized enough, that is, not secure enough), but still achieves network security.
But in daily life, it is actually difficult to define whether the number of nodes is "many or few", that is, except for Ethereum and Bitcoin, it is difficult to recognize that other networks are secure enough.
It can then be simply understood that almost all PoS networks can use Mind to achieve secure consensus.
The core logic of PoS: nodes stake tokens to obtain verification rights and participate in verification to receive rewards
This process is reflected in Depin, AI and most networks, so Mind's solution can be applied to the above types of projects.
For example, Bittensor needs to stake TAO as a validator; Io.net needs to stake IO to participate in the node, etc.
For decentralized AI, since the process involves the calculation of a large amount of high-value data, security and privacy are particularly important, and it is also necessary to participate in the calculation.
The encryption calculation process happens to be in line with the core of FHE, so FHE + decentralized AI is like Audi Double Diamond, my partner
Summarize
The basic logic of Mind:
The common problem faced by AI and networks using PoS consensus: insufficient decentralization means inherently insufficient security
FHE encrypts the voting process, forcing validators to make independent evaluations
The establishment of voting rights comes from staking, and re-staking assets can provide an additional reward compared to native assets
2. Remote Staking solves the security of pledged assets
3. Decentralized AI faces more data and security and privacy issues, such as cross-validation, model selection, training, reasoning, decentralized computing, etc. FHE perfectly solves these problems.
4. The Mind project itself is complex in design, but for users, it is actually very simple. Stake LST/LRT and get rewards
5. Mind has strong resources and background. What we need to look forward to and pay attention to now are the reward settings, token models, network support, and the first batch of Subnet applications.