Matt Brown

Interchange in 1,000 words (aka the guide I wish I had starting out in fintech) mtb.xyz/notes/intercha…

1/ One of the few "moats in the AI era" posts actually worth reading TL;DR - "Hard" fintech (regulated, built on historical performance, deals with complex licensing) just got *more* valuable in a world of abundant software From @Mattbrown - notes.mtb.xyz/p/fintechs-moa…

Full article here: x.com/Mattbrown/stat…

The test for any fintech: - Do you underwrite risk yourself or pass it to a partner? - Do you own the regulatory relationship or rent it? - Does every transaction make YOUR models sharper? - Is your ledger the source of truth?

"Fintech's moats don't compile" Everyone says AI kills software moats. But fintech isn't simply a subset of software... it just cosplayed as one for a decade to get the multiple. And AI is about to expose who was faking it.

So many inbounds for my pov on PayPal's predicament. I've been gone for 17 years, but have continued to invest heavily in payments and credit innovation. My two cents on this:

Inspired by @JayaGup10 and @ashugarg's great article on context graphs and @jaminball on evolving systems of record: some thoughts on why context graphs are so valuable, but how they're also already here... just hiding in vertical software.

That's the right frame: stablecoins vs. ACH/SWIFT, not vs. card networks. Cards also bundle consumer value (spend protection, rewards) that'll be difficult to replicate with stablecoins. No demand-side pull for checkout, just supply-side cost savings. Same reason many consumer "ACH checkout" products flopped.

Why are folks still wasting time trying to recreate a merchant acceptance network for stablecoins? It should be a foregone conclusion at this point that the killer use case for stablecoins in commerce isn’t at the point of sale; it’s behind-the-scenes settlement, after authorization is captured over card rails. For those pouring good money after bad trying to roll out stablecoin checkout — you simply cannot create a two-sided network without incentivizing all the critical nodes in the current structure to switch. And by the time you find the market-clearing rate that gets everyone bought in, you’ll notice you just reinvented the status quo. You’re fighting immutable behavioral economics by trying to end-run the card networks. The moat has hardly ever been about technology; it’s about standards and governance. Instead, the real opportunity lies in stablecoin-backed cards. Cards have ubiquitous acceptance, while stablecoins are a phenomenal way to store value across borders. The next era of payments will belong to builders who can marry the two, amplifying the daily utility of stablecoin wallets: store value anywhere with stables, spend anywhere with cards.

There are lots of cool AI products, but few feel like magic the first time you use them, and even fewer still feel like magic the 100th time. @suno is squarely in the magical category. Congrats @MikeyShulman @antrod 🥁

@tjparker Same energy as this gem

‘996’ at a laptop job has the same energy as vc’s wearing carhartt merch. Blue-collar cosplay.

Agentic interfaces really need domain specific primitives and integrations to reliably nail outcomes for non-technical users and go beyond demoware. We're going to see AI-first startups slogging to backfill those as aggressive pre-AI mid stage companies layer AI interfaces on top of years of deep engagement with customers. Parabola is a perfect example of the latter and their new AI phase is here!

@jsfranklin221 @Pocket @matter

Really bummed to see @Pocket shutting down. I’ve been a 5x+ DAU for many years. And I hate changes to my routines 😠 What else is as simple and reliable?

What if you could shop for healthcare the same way you shop for everything else? Broad selection (online and offline), clear upfront pricing with any provider (cash and insurance), and seamless booking. Introducing General Medicine -- the first comprehensive healthcare store.