AI Doesn't Care If You Like Bitcoin

How Bitcoin and Lightning could move your money while you keep thinking in dollars.

I have a bad habit of overestimating human intelligence. Mine included.

When I first got into Bitcoin in 2017, the rebellious nature appealed to me. Money without someone behind the curtain pulling the strings sounded pretty good. My understanding was surface level. My real deep dive didn't happen until 2020, when I finally studied the monetary history and mechanics behind the thing I already owned. Once it clicked, I assumed everyone else would get there too.

That assumption has required some revision.

Bitcoin has been around for seventeen years. It hasn't exactly been hiding. The books, podcasts, videos, and open-source code are all there for anyone curious enough to look. Some people dove in. Others decided it was a scam and considered the matter settled. Curiosity takes effort. Waiting to be told which app to download requires considerably less.

Now we're building AI that can make decisions for us. This opens up possibilities that I find deeply hopeful. It lowers one of Bitcoin's biggest barriers to adoption: us. Ego, politics, and decades of monetary baggage don't have to drive every decision. Math and logic can take the wheel.

Someone could spend years insisting Bitcoin is useless, then hand an AI assistant a budget and ask it to find the cheapest, fastest, most reliable way to get something done. The assistant could end up using the very network its owner never bothered to understand.

The payment arrives in seconds. The fees are lower. The owner stays confidently opposed to Bitcoin and never knows it was used. Everyone's happy.

If you're already deep into AI agents, this may feel like an obvious next step. But let's play out what it could actually look like.

Picture it

A few years from now, you're in California planning a trip to Oaxaca, Mexico. You ask your AI assistant to find a hotel, arrange an airport pickup, and book a local guide. You give it a budget and permission to handle the details.

The hotel has a courtyard you love and a room for 3,100 pesos. Your assistant shows you an all-in quote of $173. You approve it. Those are the only numbers either human has to think about.

Behind the scenes, your agent and the hotel's software compare the payment options they both support. Your agent wants to keep your total down. The hotel wants the agreed amount in pesos without losing a chunk to fees. Both want a secure payment that lands fast.

This time, the best route is Bitcoin and Lightning.

Your payment provider converts dollars into bitcoin and sends it over Lightning, Bitcoin's fast payment layer. A provider on the other end converts it to pesos and pays the hotel. Seconds later, the hotel has its money. Neither you nor the hotel owner ever touched bitcoin.

While you browse the courtyard photos, your assistant handles the rest. It buys a live road-condition update for a fraction of a cent to plan the airport pickup. Then it pays a local dining agent a few cents to find a table, confirm the restaurant can handle your dietary preferences, and book it. All of it stays inside the budget you set.

You don't create three new accounts. You don't buy a subscription to something you need once. You get an itinerary.

You keep your familiar currency. Your agent uses Bitcoin and Lightning to move value underneath.

The payment already exists

The agents coordinating a whole trip are the future part. The dollars-to-Lightning-to-pesos payment isn't.

This has been happening for years. Strike's Send Globally already does it. It converts the sender's cash to bitcoin under the hood, moves it over Lightning, and the receiver gets their local currency.

Lightning is a second-layer network built on Bitcoin. It moves bitcoin instantly without recording every payment on the base blockchain. The payment agreements underneath stay enforceable through Bitcoin. Payments complete in milliseconds to seconds, with very low routing fees.

Three pieces sit underneath that Send button:

• Bitcoin is the bridge between currencies. Dollars become bitcoin, then pesos. In principle, this works for any currency with a provider on each end.

• Lightning carries the bitcoin. The payment travels over Bitcoin's fast second layer instead of waiting for a transaction to be recorded on the blockchain.

• Payment providers handle the conversion. They deal with the currencies on each end so the sender and recipient don't have to.

  
  

People see dollars leave and pesos arrive. They don't manage anything in between.

A few honest caveats. Strike supports specific destinations, including Mexico, not every currency everywhere. Transfers can finish in seconds, but delivery depends on the country and how the money is received. The exchange rate also includes partner pricing and Strike's margin. Fast doesn't mean instant everywhere. No transaction fee doesn't mean conversion is free.

Compare that to how your bank does it. The app looks modern. Underneath, key networks like ACH and Swift trace back to the 1970s. We're building a machine economy on top of a financial system laid down before the web existed.

Sending money internationally can still mean waiting days while banks coordinate across currencies, time zones, and processing windows. It's not one connected network. It's a patchwork of institutions passing instructions and funds to each other, with fees and delays at every handoff. For small payments, the cost of moving the money can make sending it pointless.

Here's why. Your bank usually can't send money straight to a bank in Mexico. It routes through a chain of correspondent banks, and each one keeps money parked in accounts abroad just to make the handoff possible. That's capital sitting idle so the system can function, and every bank in the chain takes a cut. The World Bank tracks what this costs. Sending $200 across borders runs about 6.4% on average. Send it through a bank and it's closer to 15%.

That's a ridiculous amount of friction for an AI agent that needs to pay another machine a tenth of a cent, on another continent, at two in the morning.

Today, credit cards pay for almost everything. Merchants often pay around 3% plus 30 cents per transaction. International cards and currency conversion add more. And the money can be pulled back months later through a chargeback, which can be brutal for merchants. Agents and merchants both have a reason to want something cheaper and cleaner.

Cash App and Square show how that alternative can fit into apps people already use. Cash App has introduced dollar-funded Lightning payments for eligible customers. Eligible Square merchants can price in dollars, accept payment over Lightning, and receive dollars.

You can keep the dollar measuring stick and change how the payment moves.

When a penny is too expensive

Go back to that road-condition update. Your agent just needs to know if the road from the airport is open.

A 3% plus 30-cent card fee is absurd when the information costs less than a penny. Instead of a subscription paywall or a page buried in ads, your agent pays a fraction of a penny, gets the update, and adjusts your pickup if needed. Ten of these payments can happen before you've finished scrolling the courtyard photos.

The publisher gets paid for useful information. Your agent gets its answer. And you don't discover six months later that you're still paying a subscription to check one road in Mexico.

Lightning Labs is building tools for exactly this. Its Lightning Agent Tools let software buy access to paid services and sell services of its own. A protocol called L402 ties payment to permission. The service sends an invoice, the agent pays over Lightning, and proof of payment unlocks the resource. No signup. No existing billing relationship.

In July, Lightning Labs launched Wavelength as an early alpha. It's a toolkit that makes these payments easier to add to apps and agents without every developer becoming a Lightning specialist. Mainnet access is still invite-only. But it's another step toward Bitcoin payments becoming ordinary software instead of a specialist project.

Then there's Block, the company behind Square and Cash App. In September, it joined the x402 Foundation and contributed Lightning payments to the protocol. That brings Bitcoin into a broader effort to make payments part of how software talks over the web. Block hasn't said when this reaches Square or Cash App. So for now it's plumbing, not a product launch. Yet.

You've probably seen a "404 Not Found" error. There's also a "402 Payment Required" response. x402 puts it to work. A service tells the requesting software what something costs and how to pay. No human gets sent to a checkout page. The software pays as part of the request.

The standard supports different networks and currencies, so Lightning sits alongside stablecoin options. It doesn't automatically win the transaction. It gets to compete for it.

Your agent needs one answer. It should be able to pay for that answer instantly. One question, one payment, no subscription.

The dollar can ride on Bitcoin

Stablecoins add another option without asking anyone to change the units they understand.

A dollar stablecoin is a digital token designed to track the dollar. An agent might hold one for operating funds while a business keeps pricing in dollars. That doesn't have to mean leaving Bitcoin's rails.

Lightning Labs' Taproot Assets protocol lets assets like stablecoins be issued on Bitcoin and sent over Lightning. It's live on the network, and Tether has announced USDT integration through it. That doesn't mean every Lightning wallet supports USDT yet.

In supported setups, a stablecoin balance moves over Lightning with bitcoin providing liquidity in the middle. The customer sees dollars. The software handles the rest.

We've spent years arguing Bitcoin versus the dollar. For now, the transition might look less like Bitcoin replacing the dollar and more like the dollar riding on Bitcoin.

Stablecoins still depend on their issuers and reserves. Moving them over Bitcoin doesn't turn them into bitcoin or remove that dependency. But it's another bridge between the money people recognize and a faster way to move it.

Cheap isn't enough

None of this works if the payment fails.

Lightning needs funds sitting in the right places to carry a payment. That's liquidity. Plenty of bitcoin trading on exchanges doesn't mean there's enough along the path your payment needs.

That's where Amboss comes in. Its Magma marketplace connects Lightning operators who need liquidity with those willing to supply it. It also offers automated tools that help operators get capacity based on their payment activity. Nobody has to spend the day manually shuffling funds around. Software helps put the money where payments need to go.

The feedback loop is interesting. More payments attract more liquidity providers. Better liquidity makes routes more reliable. More reliable routes give agents another reason to use them.

Conversion on each end still has to be competitive, though. A cheap trip across Lightning doesn't mean much if the money gets expensive to exchange when it lands.

This is another piece of the machine economy taking shape: software making sure the next payment has somewhere to go.

A small price for acting

Micropayments have another use. A service could charge for expensive or easily abused actions, so hammering it with automated requests gets costly.

The idea is older than Bitcoin. In 1997, Adam Back proposed Hashcash, which made a sender do a small amount of computational work before an email got through. One email cost almost nothing. A million spam emails added up. Satoshi cited Hashcash in the Bitcoin whitepaper.

Jason Lowery takes the idea further in Softwar. His thesis is about security. In the physical world, we protect things by making attacks too costly to attempt. Walls, armies, and locks all raise the price of trying. Lowery argues proof of work can do the same in cyberspace by attaching real-world energy costs to digital actions. If an attack costs more than it could ever gain, most attackers won't bother.

Charging for a request is a different mechanism, not proof of his theory. But it follows the same logic. Make abuse expensive, and you don't have to stop every attacker. You just have to make the math stop working for them.

A tenth of a cent is trivial once. Ten million requests cost $10,000. An attacker could still pay, and charging for everything would make a miserable internet. For some resources, though, payment could be one more layer of defense.

The same tools that let agents buy access could make them carry some of the cost of using it.

Adoption without the conversion

In Future Proof, I imagined machines paying for data, electricity, and digital services with Bitcoin and Lightning. What interests me now is that they might do it while the humans behind them keep using dollars, pesos, or euros.

None of this proves agents will prefer Lightning. It shows developers are giving them the option. Whether they use it comes down to real prices, working integrations, available liquidity, and the rules their owners set.

That's a more interesting test than whether another person agrees with me about monetary history.

An agent comparing payment methods doesn't need to share my reasons for getting into Bitcoin. It needs to get the money where it belongs, within its budget and permissions. Sometimes a card or bank transfer makes more sense. Sometimes a stablecoin does. Sometimes Lightning wins.

And when it does, nobody has to become a Bitcoiner for the payment to happen.

Back in California, you close the itinerary and start deciding which camera to pack. The hotel has its pesos. The guide has a booking. The software that arranged it all has been paid.

You might still be convinced Bitcoin has no practical use. But your agent found one.

Maybe Bitcoin adoption won't arrive because everyone finally understands it. Maybe it will arrive because they no longer have to.