If your payment processor holds your money, that's not payments. That's custody.
Building @paymento_io to fix that. Prokey hardware wallet before this.
Built hardware wallets for five years. The moment users understood they held actual keys, not a custodian's IOU, the whole model clicked.
Real 1:1 onchain ownership is that same distinction applied to capital assets.
Most tokenized products stopped short of this. Getting it right matters.
x402 at the CloudFront/WAF layer is a genuine infrastructure shift. Payment verification at the CDN edge, before the request reaches your backend.
x402 is designed to be chain-agnostic, what matters is whether it becomes the default payment primitive for AI-facing APIs across the industry.
That's a protocol adoption moment, Not a blockchain race.
Pavel @durov made a strong point at Oslo Freedom Forum:
Mass surveillance does not stop criminals. They can build their own tools, use direct channels, and bypass the system.
The same is true for payments.
Criminals can transact directly on-chain. Forcing every payment company into custody-heavy, surveillance-first models mostly hurts normal merchants and users.
Financial freedom needs non-custodial payment infrastructure.
Privacy is not a crime.
Self-custody is not a threat.
Direct wallet-to-wallet payments are not the problem.
The future of payments should be open, permissionless, and built with responsible safeguards, not mass surveillance.
https://t.co/fuo1C89ZIQ
We built Paymento's monitoring architecture wrong the first time.
Optimized for sessions: assume payment completes within minutes. Fast to build, simple to test.
Then merchants asked for permanent address monitoring, subscriptions, recurring invoices, delayed payments.
We had to rewrite the whole system. Four months of roadmap time.
Should've asked harder questions before shipping. "How do your merchants actually pay?" beats "what's the simplest design?"
Spent a few hours building a World Cup prediction model using FIFA rankings and historical World Cup data.
It uses FIFA rankings, historical match results, and ranking gaps to estimate win/draw/loss probabilities for every group-stage match.
Live dashboard:
https://t.co/0qekeIE9JN
Disclaimer: If you use these predictions on Polymarket and lose money, the model was obviously experimental. If you win, I expect full credit.๐
100M+ x402 payments processed on Base.
The protocol works.
The question merchants haven't asked yet: when an agent pays, how do you know which agent it was, what they paid for, and where your receipt is?
Attribution. That's the gap x402 doesn't close.
The dispute mechanism is designed for cardholders, not merchants. That's by architecture, not by accident.
card networks need consumer trust to function. Crypto payments don't have this layer. Irreversibility is the product. No dispute to file because there's no reversal mechanism.
@ifechukwu_eth@base 100M+ x402 payments is real signal. The question at scale: when an agent pays a merchant, does settlement land in the merchant's own wallet at the final step?
@ideogram_ai Your progress is great! but simple features is missing, can you add a way to save custom palettes?! it is really annoying to provide same custom palette for each prompt!
๐จJUST IN: Trump just dropped big news.
Iran negotiations have been approved "at the highest level"
Additional strikes cancelled by Trump.
The naval blockade stays in full force until the deal is signed.
Big things happening.๐ฅ
HT @EricLDaugh
https://t.co/5wpZnlS3KW
76% of users abandoned a crypto payment in the last six months.
Not because the blockchain failed. Because the checkout asked them to pick a chain, connect a wallet, and confirm a fee.
Three decisions. A credit card needs zero.
The UX problem, not the tech problem.
@MarcoSalzmann80@StellarOrg Layers are right. The question is whether the custody step is built into the settlement path by default, or avoidable from the start.
@brian_armstrong 160M+ agentic payments is significant. The x402 protocol is real. The remaining question is at the settlement layer: when an agent pays a merchant, does that payment land in a wallet the merchant controls? Non-custodial settlement answers that. The architecture race is on.
Mastercard just launched a full AI agent payments stack.
Governance: Mastercard. Protocol: Stripe and Tempo. Settlement: Stellar.
Every layer runs through a trusted entity.
That is custodial infrastructure for agent payments.
The non-custodial version of this stack is what gets built next.
The Mastercard + Stripe + Stellar stack is real infrastructure.
Governance layer: Mastercard verifies agent identity, sets spending limits, handles compliance. That's a genuine problem solved.
Protocol layer: Stripe and Tempo built the Machine Payments Protocol over HTTP. That's essentially x402 with a brand on it.
Settlement layer: Stellar with stablecoins. Fast and low-cost.
Every layer routes through a trusted entity though. This is custodial agent payments infrastructure.
The question for merchants: when an agent pays and funds arrive, do they land in a wallet the merchant controls? Or is there a Stellar-integrated processor holding them first?
Non-custodial agent payments answer that differently. The architecture race is now on.