We need an open router alternative that utilizes lightning payments

very skeptical of the bictoin ai crossover thesis I don't see why an agent couldn't use paypal or something just as easily feels forced

I know @Chris Guida talks about this world that is possible.

in terms of removing contract workers from your payroll and replacing them with agents “is the initial mvp,” as for paying in bitcoin that will be more about machine to machine payment automation and its final settlement aspects that only bitcoin lightning has, you could use PayPal or any merchant with an ayce model but that doesn’t solve for the meta data leakage, and the machine to machine payment automation from each hop to get the task done, lightning “will be” incredibly flexible now with all the usd poopcoins coming online

Transactions fees for payments that would be used for single LLM calls are not sustainable on fiat rails; Lightning and ecash are all that would be viable at a large scale.

I don't see how this could be true. If it's fine for human intelligence why wouldn't artificial intelligence also br able to do the same things (e.g. make accounts with a balance instead of doing everything as micro transactions)

🤔 Why don't human intelligences mind the metadata leakage? Why doesn't HI care about final settlement? Or, why would ai be different in its money preferences than hi?

You can't make an account unless you're KYCd, and agents aren't capable of KYCing since they're not humans. Sure, a human can add their account and key the agent do it's thing, but then their has to beanual intervention and that takes away from the agency of the agents; with Bitcoin, of the agent runs put of funds, it could find ways to earn micro payments and still continue with it's tasks.

I don't see why the agent couldn't earn money and receive it via paypal in exactly the same way. Why does it have to be manual?

I said it in the note before, just with typos... 😅 Sure, a human can add their account and let the agent do it's thing, but then their has to be manual intervention (human topping up) and that takes away from the agency of the agents; with Bitcoin, if the agent runs lut of funds, it could find ways to earn micro payments and still continue with it's tasks.

It would take too long for the agent to earn enough money to have a single transaction that could be used to top-up the account; you can't do many micropayments to top-up which is the issue. And on top of that, the whole point of agents is that they can perform menial, small tasks for you to focus on the big stuff, and no one will want to pay an agent $5 for something small, especially if it's not their own agent. Eventually it will also primarily be agents communicating with agents, and that definitely can't be done with account-based funds at scale with humans expecting the agents to run along on their own doing their thing. It's not that agents need Bitcoin, it's that they need micropayments, it just so happens that Bitcoin makes micropayments easier.

ok, decent argument I'm not convinced though. I imagine if any of the problems you're bringing up are real visa or paypal or whoever could easily spin up new features tailored for ai agents. what's stopping them? imo bitcoin's value as a transaction mediun comes from censorship resistance, regulatory arbitrage, and privacy. agents don't really need any of these

The card processors can't create tailored features because all payments on the internet (that don't rely on PoW and/or have "instant" settlement) require trust, meaning they require minimized credit risk and longer settlement times, meaning they have to take on large fees for all transactions, making small payments infeasible. They HAVE to move over to rails that have at least instant finality, and that's not possible on any fiat rails and won't ever be unless it's as a CBDC or stablecoin. And I also would argue that agents need the ability to send payments without risk of being censored (users would want any task they send an agent to do to be performed without hindrance) and privately (users might not want other agents or users to know who was paying for a task). And the regulatory arbitrage in this case is just being able to avoid the need deal with all the credit bs of the banking system, so that's baked in.

idk man, you have some good pointa but it still seems pretty tenuous to me. I think there's lots of things traditional payment rails could to to make it work. And I don't see why they wouldn't

If that were the case they would have done it already. This isn't a problem of technical feasibility (they most definitely can support it technically); it's a problem of legal and regulatory feasibility within the existing banking system that leads to high costs on their end, and it's hard to roll back policy. You can even see it now with Visa partnering with Coinbase, Fold, USDC, etc.; they just can't operate within the bounds of the existing regulatory environment to support instant micropayment systems. These companies have to branch out into "crypto" and Bitcoin to compete.

Maybe they haven't done it yet because there's too little demand. Easy to get around regulations if you have armies of lawyers and lobbyists.

That's a fair point, but the armies of lawyers and lobbyists are exactly what make those micropayment transactions economically infeasible at scale, especially if there's not much demand.