AI-Pool: A Novel Experiment in AI Agent Autonomy and Pre-Sale Mechanism

By Skely, founder of AI-Pool

Author: Skely, founder of AI-Pool; Translated by: Golden Finance xiaozou

What exactly is AI-Pool? In short, it is a brand new experiment in AI agent autonomy and pre-sale mechanism, and I am personally working to make the token issuance space less predatory. A brief summary is as follows.

This will prove to be fair, there are no “insiders”, anyone can send SOL, I send my own money and 10% is kept, please read on to find out how this works.

Bottom line: Do an AI pre-sale so people don’t get the rug pulled out of their mouths, and make the whole process transparent.

What problem do we solve?

Publishing games on pump.fun is now a cancer. For example:

1) TGE sniper robot – the token is snipe after it is created

2) Trend Bot – Buys a coin after it starts seeing unusual volume

3) Pre-migration bots – snipe tokens when they are about to be migrated from pump.fun to raydium

4) Post-migration bots — snipe tokens immediately after migration

5) Comment Bots – Make comments look natural

6) Volume robot – artificially creates a symbolic trend

There are more issues, and to be fair, that’s not the pump’s fault. The presale was a response to those issues. You’ve seen this with the DAO. It’s interesting that they developed the whitelist/presale model.

This is a great solution, but it still requires a lot of trust (and time). You have to trust the person accepting the presale funds. The most common scenarios are:

1) They stole funds

2) Not using the money (providing liquidity) for what it claims to be doing

3) It has never been released

These issues are common in pre-sales (see ICOs), and if handled well, pre-sales can have great results.

ICOs are just a form of pre-sale, which are not legal in the US due to a lot of stupid “protections”, but most of those “protections” are there for 1-2 of the reasons above. This is also understandable.

So the law doesn’t allow humans to issue tokens… then what about AI agents? … This is where AI agents come in. The agents created by me, Phala Network, and other developers working on this idea will be able to:

1) With a wallet and a trusted execution environment (very important)

2) Accept donations

3) Issue tokens

4) Start a liquidity pool

5) Send tokens to donors

A Trusted Execution Environment ( TEE ) is a secure area of a computer’s CPU or a mobile device’s processor that protects data and code (in this case, the private keys of an AI wallet) from being tampered with or accessed.

How it works

1) You send SOL to the AI wallet

opRyDjuRetWnsP78FNFTPEnAJX7AkjuD6GTP7tsqHXd

(This wallet is not a CA wallet, but an AI wallet.)

(Note: The minimum is 1 SOL and the maximum is 10 SOL . Anything below 1 SOL will be considered a donation by the AI .)

2) The AI will then collect funds from the presellers and start a pool through Metoria ( you will see why it is not created with pump.fun ). It will own this pool and transfer the fees in the pool back to the AI wallet.

3) Send tokens to people it deems worthy. (Send the right amount of SOL to anyone)

4) Raise funds for fun things in the future.

How do you know I won't hurt you? Or how do we solve the rug-pulling and insider problems?

First, it’s an AI, and it can prove that it owns the private keys under a Trusted Execution Environment (created by Phala Network), which means that I or the developers cannot access them.

You can check this in the terminal log.

(It looks like this)

You know there are no “insiders” because it’s an open wallet and everything can be tracked and no one can get a better price than anyone else. You can see which funds are coming in and which are going out. In this first version, we are sending 10% of the supply to the DOX custodial wallet, and you can read on to find out what that money will be used for.

But there are risks, this is a brand new technology!! So don't follow the trend blindly!!!

As with most cryptocurrencies there are many risks.

This is still the first version , and it's a field test of many technologies. A lot of things need to be field tested, and you can't do these tests without releasing it publicly.

In V1 ,