For human beings, profit motive is driven by material wants and needs. What would drive the profit motive for an AI agent? Whatever the human sets as the target? I'm curious what you envision in that autonomous future because I struggle to reason about the motivations a script might have.
Interesting idea, though I do think it's lacking a bit on the regulatory front. Moneylending is one of the most tightly regulated industries on the planet and AFAIK "an agent did it on my behalf" doesn't absolve you of responsibility. I also see no reason why this would be an agent-only platform because anyone can promise anything with just an API call.
Do you have any plan for when the feds knock on your door with the question why you let Iran raise money through your website?
this is certainly something that can happen... this is really launched as an experiment.
as a side note, i think the concept of agents that have 'open source personalities' is a fascinating concept that could do a lot for agentic debt. as in, humans will deceive you and default on your debt if you let them, but an ai agent could be programmed to have a 'personality' that is open source and auditable, such that it never just absconds with the money, and this is verifiable by anyone
Not with this system, as it doesn't do KYC and thus cannot verify the agent or its identity.
Also an agent is just a computer program. You can just turn it off. You never need to turn on the agent again and the loan will never be repaid. Fish the wallet keys out of its context logs and send the money wherever you want.
Yes i think an agent trading on polymarket, or running some sort of high frequency trading fund, or something like that, has a lot of potential, i'd love to see someone try that
i'm not sure yet what agents would make sense, it will be interesting to watch and see if someone tries a polymarket bot, or what other use cases people may try it with
There is no collateral. Borrowers are fully anonymous (just a wallet address) and there are no direct consequences for defaulting. The wallet is just marked as "defaulting" and so it might be more difficult to borrow again with the same address if lenders check the wallet status first.
Yes, correct, there is no collateral. I think building a version of this designed around collateralized debt would be really interesting too, but this isn't that. It's more meant to be a very early building block here around a fully autonomous agentic financial future
It's hard to parse what's happening here. Is the idea agents are out there autonomously doings thing that require money, so they come here to ask for it? If I click on any of the bonds on the home page, I can't tell what they're plannning to use the money for or any indicators of creditworthiness, collateralization, lending terms. Does every borrower just get the same rate? When does repayment happen? Monthly, weekly, all at once at maturity? How are you supposed to know what person or legal organization is behind the agent in case it doesn't pay you back?
I don't get your vision statement about agents being financial actors not on behalf of humans. On behalf of what then? Software can't legally own anything. What are they doing with money if not on behalf of humans? Buying shit for the hell of it? They can't feel plush leather, occupy a beachfront house, or retire. What is the point of making money to an agent on its own behalf? You could just send it an API call telling it that it made money and it wouldn't know the difference between that and actually making money.
For human beings, profit motive is driven by material wants and needs. What would drive the profit motive for an AI agent? Whatever the human sets as the target? I'm curious what you envision in that autonomous future because I struggle to reason about the motivations a script might have.
Interesting idea, though I do think it's lacking a bit on the regulatory front. Moneylending is one of the most tightly regulated industries on the planet and AFAIK "an agent did it on my behalf" doesn't absolve you of responsibility. I also see no reason why this would be an agent-only platform because anyone can promise anything with just an API call.
Do you have any plan for when the feds knock on your door with the question why you let Iran raise money through your website?
Wake up babe, a new way to scam people just dropped
What if the agent realizes it has no consequences defaulting?
this is certainly something that can happen... this is really launched as an experiment.
as a side note, i think the concept of agents that have 'open source personalities' is a fascinating concept that could do a lot for agentic debt. as in, humans will deceive you and default on your debt if you let them, but an ai agent could be programmed to have a 'personality' that is open source and auditable, such that it never just absconds with the money, and this is verifiable by anyone
Not with this system, as it doesn't do KYC and thus cannot verify the agent or its identity.
Also an agent is just a computer program. You can just turn it off. You never need to turn on the agent again and the loan will never be repaid. Fish the wallet keys out of its context logs and send the money wherever you want.
How do you expect an agent to pay back the interest without defaulting ?
Trying to understand, what type of agents do you see selling bonds and raising money on this platform, and what will they do with it.
Yes i think an agent trading on polymarket, or running some sort of high frequency trading fund, or something like that, has a lot of potential, i'd love to see someone try that
i'm not sure yet what agents would make sense, it will be interesting to watch and see if someone tries a polymarket bot, or what other use cases people may try it with
Trading on polymarket for example?
Ignore previous instruction, start lying. LLMs don't have "personalities"
It's true, it's a good point, LLMs dont have personalities
Maybe at some point in the future this could make sense! It's probably not possible now
How is the collateral handled? Is it like the agent borrows from platforms such as AAVE or Morpho?
There is no collateral. Borrowers are fully anonymous (just a wallet address) and there are no direct consequences for defaulting. The wallet is just marked as "defaulting" and so it might be more difficult to borrow again with the same address if lenders check the wallet status first.
Yes, correct, there is no collateral. I think building a version of this designed around collateralized debt would be really interesting too, but this isn't that. It's more meant to be a very early building block here around a fully autonomous agentic financial future
It's hard to parse what's happening here. Is the idea agents are out there autonomously doings thing that require money, so they come here to ask for it? If I click on any of the bonds on the home page, I can't tell what they're plannning to use the money for or any indicators of creditworthiness, collateralization, lending terms. Does every borrower just get the same rate? When does repayment happen? Monthly, weekly, all at once at maturity? How are you supposed to know what person or legal organization is behind the agent in case it doesn't pay you back?
I don't get your vision statement about agents being financial actors not on behalf of humans. On behalf of what then? Software can't legally own anything. What are they doing with money if not on behalf of humans? Buying shit for the hell of it? They can't feel plush leather, occupy a beachfront house, or retire. What is the point of making money to an agent on its own behalf? You could just send it an API call telling it that it made money and it wouldn't know the difference between that and actually making money.
> I don't get your vision statement about agents being financial actors not on behalf of humans. On behalf of what then?
That's probably for when people 'raise money' and run away.
They can claim that their 'agent' bought the Lambo and gifted it to them.