jackie | agaperste

Issuance @Stablecoin (Stripe) 🧙 new money meets fiat rails | prev @UniswapFND @Dune @Fidelity | Tinker about human existence ✨🤖⌛🔁 | @UnordiJourney 🎙️

[jackie | agaperste @agaperste]

Sometimes I forget and think of faster redemption as the “mundane” stablecoin use case. Then I get reminded that’s not true when I hear it’s the actual reason developers pick stablecoin rails, or pick one stablecoin over another (with some stable, free 1:1 redemption can mean waiting up to 2 business days).

The simplest use case still makes me giddy, especially when I see it play out with OUSD at @Stablecoin.

[jackie | agaperste @agaperste]

Good data design and good payments design follow the same principles: be trusted, stay invisible, and stay compatible both backward and forward.

BNY settles around $2.5 trillion a day and most people have never thought about them once. A data model is the same. If people are spending time thinking about your schema, something's wrong.

Compatibility works the same way too. Stablecoins don't win by ripping out fiat rails, they win by plugging into them so nothing breaks for people already moving money. Same with a data warehouse: built to handle whatever comes next without breaking anything downstream.

While thinking about this it just dawned on me that this is why I'm so excited about what I'm doing at @Stablecoin and Stripe. We're wholesale transforming payments, and I get to be part of making that happen. Good feeling.

[jackie | agaperste @agaperste]

Spent lunch today watching a few product demos (BVNK, Zero Hash, DFNS etc.) and the thing that stuck with me isn’t any feature necessarily. It’s that payments is a trust business, and trust, in colloquial terms, is vibes.

Kinda joking, but not really. When you get MTLs or pursue OCC, that mountain of application materials exists to objectively show your vibes are trustworthy: the controls are real, sanctions get screened, non-compliant payments get stopped, PII stays protected even internally. Unglamorous operations, but that’s where the trust actually lives.

Once you’ve earned it, the goal flips: you want to disappear. Everyone knows Visa, yet nobody thinks about Visa while paying. You become visible again when there’s friction. Take SWIFT, for example. Not a knock on SWIFT at all, it’s the superior version of cross-border, and any payments company that unlocks it is glad to have it. But even a rail that good can be a friction point, because payments still take days, and days are long enough to notice. That’s what makes the stablecoin rails interesting to me: they could make the noticing go away.

For data work, I keep going back and forth on whether the same rule applies. I don’t think the dream state is literally zero friction or full invisibility. Maybe it’s invisible visibility: data models designed so well that analytics are self-serve, operations always have what they need, and sometimes an insight surfaces unprompted, pushed by an agent loop you built, right when the business needs it. Enhanced eyes and ears for the business, without the business waiting for you to do it.

[jackie | agaperste @agaperste]

Lunch cooking thought: the best data analysts are secretly journalists.

Same job, really. Dig through a mess of information, find the story that matters, tell it so ppl act on it. It's just most do it internally while the select few do it publicly.

[jackie | agaperste @agaperste]

@MSilb7 Hah

[jackie | agaperste @agaperste]

Why do developer accounts so often sit under a legal entity name that looks nothing like the brand? I look at this data every day and it still trips me up sometimes. It's already a lot to hold developer names in my tiny head with limited RAM, and now we gotta double the number of names? 😂🌝 jk it's a good problem to have I guess

[jackie | agaperste @agaperste]

@charliekerr @andromeda_ai congrats!!

[jackie | agaperste @agaperste]

AI makes you a better thinker, if you apply it correctly. It took the friction out of my execution and moved it into my thinking. 🤖🙆

Two things from this week. I was getting a collateralization reporting pipeline right, and AI wrote the SQL faster than I ever could. But it had no idea what “right” meant. Which balances count, which edge cases are real, what the number is for. That lived in my head, and the work was getting it out clearly enough for the tool to move.

Then I was doing company and market research for Econ Labs, and prompted well, it did something different. It pushed back, opened angles I hadn’t considered, made me doubt my own framing. In a good way.

Every time I skipped the thinking and just handed something over, I wasted tokens, and more importantly time and energy. With AI, direction matters more, not less.

Which is why a line from Eileen Gu’s SMU interview in Singapore stuck. She quoted the part before Descartes’ famous one: dubito, ergo cogito, ergo sum. I doubt, therefore I think, therefore I am. There’s more surface now and the work moves faster, so you have to doubt more. Most days that means thinking carefully, picking one direction on purpose, then executing relentlessly without getting pulled into reels or a random Slack message 🥹

So what stays? Curiosity, I think. Being in your body. Art, actually.

[jackie | agaperste @agaperste]

@W3Wag3s @casaisdev @Stablecoin Precisely, we are not longer solving for tech issues alone, it’s more the everything else after technology

[jackie | agaperste @agaperste]

A friend interviewing with us told me the problem we gave him felt weirdly elementary.

I laughed because I remember thinking the same thing my first week at @Stablecoin. The first thing I shipped was a reference table. A table! Mapping some things to some other things. I was a little embarrassed to even bring it up in meeting.

That table is now underneath a big chunk of how we monitor and operate the tokens we issue. I’ve rebuilt it more times than I care to admit, not because the idea changed but because it had to get more scalable as we launched more tokens and more automated as data surface area got bigger. Nothing about it is clever. All of it has to be right.

That’s kind of the whole thing with stablecoin issuance. The surface looks elementary. Underneath you’re re-plumbing how dollars move for the internet economy, and eventually for how the world pays and allocates capital. At that scale the elementary stuff is the foundation, and if it doesn’t hold, none of the interesting product work on top of it happens.

Reserves are the same story one level up. Tokenization is accepted as the frontier now, and it is. But a GENIUS compliant stablecoin still can’t fully rebalance its reserves on after banking hours. The chain runs 24/7. T-bills, money market funds and Fedwire don’t yet. So “mint and redeem freely” at scale really means running a 24/7 asset on top of a 9 to 5 reserve. It’s solvable, we along with a lot of other people are solving it, and each piece that gets solved turns a conference talk into something you can ship. Payroll that lands on a weekend and parks itself to earn yield until you spend it. Treasury that doesn’t sit idle because of a calendar.

I’m writing this the same week Ramp raised $1.85B at a $60B valuation for corporate cards and expense software, $1T+ came off AI and chip stocks, and Jump and DRW are busy turning GPU compute into a tradable commodity. Both sides are exciting and both have real money in them. AI swings harder because it’s newer and nobody has settled on how to value it. Payments swings less because the model is proven and the problems are known. Known isn’t easy though, and it definitely isn’t done.

So yeah, the problem looked elementary. Most of ours do, right up until you add nine zeros and the table design or the settlement window becomes the pillar underneath the day to day product. Nobody warns you about that part, and honestly it’s the part I’ve grown to like.

[jackie | agaperste @agaperste]

@casaisdev @Stablecoin I do think more and more ppl are discussing and designing this, starting with bringing MMF onchain the right one, together with legal accounting recon etc partners

[jackie | agaperste @agaperste]

fun week to be in this industry 👀

1/ @deel went live on @tempo private zones for contractor payouts (w/ @Stablecoin issued DeelUSD!)
2/ DRW and Marex ran a total return swap on the S&P iBoxx IG index onchain (@CantonNetwork, @KaikoData tokenized the index)

one is the world I work in day to day, stablecoins and payments. the other I had to read up on 🤓 but it's the same story: nothing new financially, just the same money movement and the same instruments on better rails. both sides share one record instead of reconciling two books after

payments came first because moving money is simpler. instruments take longer, and now they're showing up one by one

honestly, amid all the selloff noise and the AI everything, this is what energizes me most. we're renovating the rails, with agents and new tooling helping along the way, and it actually means something

[Sector📊 @0xSector]

Looks like OUSD just entered the DeFi lending market.

$25m was just deposited into @SentoraHQ's new Morpho vault on @tempo. Lending APY around 3.6% atm

[jackie | agaperste @agaperste]

@Stablecoin had so much fun jamming with fellow panelists! excited to build the future together in this new design space of agents x stablecoin rails

[Bridge @Stablecoin]

An agentic stablecoin transaction may be instant and final—but when it’s one leg of a flow that moves dollars into stablecoins and back, the payment may not be fully settled.

Watch @Stablecoin’s Jackie Zhang (@agaperste) and other panelists unpack the reconciliation, tax, and regulatory infrastructure agentic payments require.

[jackie | agaperste @agaperste]

What is going on with XRP I’m so confused

[jackie | agaperste @agaperste]

Excited to jam about machines paying machines 🤖 and how AI agents, MCP, et al are reshaping payments, commerce, and ofc human behavior along the way!

last panel of the day, come say hi!

[Stablecoin Summit @stblcnsummit]

Jackie Zhang (@agaperste), Issuance & Crypto Data Lead at Bridge (@Stablecoin), will speak at Stablecoin Summit Singapore on the panel: "Agentic Payments: Infrastructure for Machine-to-Machine Settlement".

https://t.co/EESpc1VfT8

[jackie | agaperste @agaperste]

hard to keep track of the list of %agentic%protocol s

pretty cool! slightly terrified :')

[Simon Taylor @sytaylor]

Huge!

Meta and Sierra just published a standard for how your personal AI agent logs in to a business. Meta, Instinct, Stripe, Shopify and Walmart signed on.

This is some of the biggest distribution platforms in agents AND e-commerce behind one protocol.

---

Sierra (Bret Taylor and Clay Bavor) and Meta announced Personal Agent Protocol today with Genesys, Instinct, Rocket, Shopify, Stripe and Walmart. The v0.1 spec lands later this month.

Today a personal agent uses your website the way a human does. It clicks through pages, fills in forms, and phones support when it gets stuck. It's slow and it often fails.

Under Personal Agent Protocol (PAP):

- The agent reads your website to find what you offer and how to connect.
- It starts as a guest or signs in to the customer's account.
- The customer chooses read-only or write access. The session runs on OAuth.
- You choose the door: your website, your APIs (MCP, OpenAPI), or your own agent.

---

Amazon is blocking agents.

PAP is an alternative to welcome them through the front door, albeit making them go through the security line.

If you run an e-commerce website or online brand, your customer's agent arrives at your login page either way. PAP lets you recognize it, give it read-only by default, and route anything that moves money to your own agent.

Agentic everything feels more pragmatic by the day.

[jackie | agaperste @agaperste]

settlement isn't just a rail, it's a business model. in the agentic age the value in the stack moves.

my guess is it moves toward recon and reporting, which matter way more when agents are doing millions of txns
disputes and fraud still need to exist. we should co-design them with devs and end users, alongside the card networks and PSPs that already do this well, plus stablecoin infra (chains, issuers, etc)

(crypto data person now in stablecoins, still learning cards / settlement / PSPs, so take this as a learner's take)