good news: @cashapp tags are back in stock and here to stay 🪄 💳 🖤
more good news: starting today kids as young as 6 can get one. now the whole family's in on the fun
nobody is connecting the dots on what just happened on X.
grok 4.6. grok bot. cursor integration. supergrok heavy. all in one month. frontier model, agent platform, coding tool, premium tier. 30 days.
btw grok 4.6 is absurdly fast. like noticeably faster than anything else i've used. it doesn't endlessly ponder. it just answers.
i've been running grok bot every day. it logs into my actual apps and does work. not a chatbot. a coworker.
X right now: social, AI, agents, search, code, voice, images, video. payments next.
everyone laughed at the everything app thing 2 years ago. it's shipping weekly now and most people haven't noticed.
grok was mid 12 months ago. now it's the fastest reasoning model out. came from behind, never stopped shipping.
who else is doing this across this many surfaces at this speed. serious question.
Hey Omid, author here. Thanks for reading the piece and for the feedback.
A few things we think you're missing:
On interchange - there would be little incentive for issuers to issue cards if they couldn't monetize it. Yes they work where interchange is capped, but capped does not mean zero. Capped interchange just means less rewards, as you note. In places with capped interchange you also tend to see higher annual fees to compensate.
On settlement - you're right that the largest banks don't have this problem. But our post is not addressing the problems of JPMorgan. The problems we describe are very real for small and mid-sized banks and fintechs. And the rails you point to (same-day ACH, RTP, FedNow) are US systems. Much of the market issuing cards sits outside the US with no access to them at all. As a few people have already pointed out in your replies.
On cross-border - Visa settles in about 25 currencies and the Colombian peso is not one of them, so the example stands. The problems of interfacing with the correspondent banking system are real for programs in LatAm and elsewhere looking to serve a broader audience.
On where this ends up, we agree: stablecoin cards are a bridge step to mass adoption of stablecoins, which will also have merchants accepting stablecoins directly. It is hard to see how this will be done in a pure peer-to-peer manner, as people will still want chargebacks and fraud disputes.
But we will see networks evolve, or new networks emerge that operate solely in stablecoins.
Excellent article walking through how cards work behind the scenes.
Have worked on building products across card flows for five years now- first from the issuer processor side (@lithic, Rain’s processor) and now the issuer side @mercury!
Your card works on weekends. The money behind it doesn't.
When you tap, the network moves a message. The money follows days later.
Why that gap exists, and how stablecoins close it.
Rain will be one of the most defining companies of our time.
They’re quietly becoming one of the most innovative companies not just in crypto, but in the entire financial world.
Why? They’re solving the real bottlenecks that traditional networks still carry.