I wrote about MCX fifty times.
Fifty posts about a payments consortium most people have never heard of. Forty-four more about CurrentC, the app it was building. Sixty-nine about beacons, sixty about iBeacon specifically. Sixty-five about Apple Pay. Between 2013 and 2019 I published nearly a thousand posts on this site, and a large share of them were about how you were going to pay for things in the future.
Almost none of it happened.
I'm not bringing that up to be clever. I believed a good deal of it at the time, and the record is right here — you can scroll back and check. I'm bringing it up because I've now watched one technology cycle run the whole distance, from the launch announcements to the quiet obituaries, and I wrote it down as it went. That turns out to be a useful thing to have done, because another one is happening right now and it rhymes.
What everybody knew
In 2014, here is what everybody knew.
The physical wallet was finished. Within five years you would pay for everything with your phone. Retailers would know you had walked into the store before you reached the second aisle, because small Bluetooth transmitters — beacons — would be stuck to the shelves, talking to an app in your pocket. Your coupons would find you. The checkout line would dissolve.
This wasn't a fringe position. It was the consensus of an entire industry. Forrester published on it. Business Insider published on it. I published on it, constantly.
And the biggest retailers in America were not going to let Apple own it. So Walmart, Target, Best Buy, CVS, Sears, Kohl's and about fifty others formed a consortium — the Merchant Customer Exchange — and built their own payment app called CurrentC. It used QR codes instead of the NFC chip in your phone. It drew money straight from your bank account instead of a card, which cut the card networks out of the transaction and saved the merchants the interchange fee.
Some of the participating retailers signed exclusivity terms and switched off the NFC readers already sitting on their counters. You could not use Apple Pay at those stores. That was the point.
Meanwhile, the wireless carriers had their own entry — ISIS, which had to be renamed Softcard for reasons that became obvious in 2014.
Enormous companies. Enormous budgets. A genuine strategic rationale. Years of runway.
What actually happened
Softcard was bought for its patents and shut down inside of a year.
CurrentC never made it out of a pilot. MCX — the combined weight of the American retail industry — wound down without shipping a product at national scale. Paydiant, the company that built its technology, got absorbed by PayPal.
The beacons mostly came off the shelves. Walk into a store today and count them.
Apple Pay survived. So did the other phone wallets. But not in the form anyone described in 2014 — they became a slightly faster way to use the same credit card you already had, running on the same card networks that MCX was formed to escape. The revolution shipped as a convenience feature.
And here's the part that took me longest to see: the failures were not total either. Walmart shipped Walmart Pay in 2016. It uses QR codes, runs through the merchant's own app, and it works. The proximity technology behind beacons didn't disappear; it dissolved into store apps and order-ahead pickup, where nobody calls it a beacon and nobody writes think pieces about it.
Nothing was transformed. Nothing collapsed. Everything got absorbed, quietly, in a smaller and less interesting shape than either side predicted.
Three things I'd tell my 2014 self
**One: ask whose problem it solves, and then ask who is being asked to adopt it.**
CurrentC existed because merchants wanted to stop paying interchange fees. That is a real problem — for merchants. It was then packaged as a consumer product and handed to shoppers who were asked to link their checking account to a QR-code app in exchange for approximately nothing.
Apple Pay solved a customer's problem, badly and marginally, but it solved a customer's problem. That was the whole difference. Not the technology. Not the money.
When the party that benefits and the party that has to adopt are different people, the adoption doesn't happen — no matter how much capital is behind it.
I would now ask that question of every enterprise AI deployment I read about. A tool sold as productivity for the worker, purchased by an executive who has been shown a headcount projection, is not a productivity tool. It's a CurrentC. The person asked to adopt it can generally tell.
**Two: a consortium is a tell.**
I read MCX as a show of force. The largest retailers in the country, aligned. How do you beat that?
I had it backwards. Companies that are winning don't form consortia. They ship. A consortium is what happens when a group of incumbents is frightened of the same thing and would rather coordinate a defense than compete. It's a lagging indicator of fear, and it moves at the speed of its slowest committee.
I now read industry alliances, joint standards bodies, and shared-principles announcements in that light. Sometimes they're substantive. More often they're the sound of large organizations trying to slow something down.
**Three: the honest forecast is boring, and nobody rewards it.**
The two loud positions in 2014 were "this changes everything" and "this is a bubble." Both were wrong, and both were more fun to write than what actually occurred, which was: *some of this gets absorbed into ordinary practice over about eight years, in a form that isn't very exciting, and the interesting question turns out to be who captures the money.*
Nobody clicks that. I know, because I mostly didn't write it.
What I got wrong
I want to be specific, because a retrospective where the author turns out to have been right is worthless.
I gave beacons far more coverage than they earned. Sixty-nine posts. I was persuaded by the demos, and demos are designed to persuade. I did not ask often enough what a shopper actually gained from being detected.
I treated the MCX-versus-Apple story as the main event for years. It was a sideshow. The main event was that payments are a network business, and the networks were never in danger from any of it.
And I mistook activity for progress. Pilots, partnerships, press releases, conference keynotes — I covered all of it as though motion were the same as traction. A great deal of what I wrote is a record of an industry talking to itself.
So what does this say about AI
I want to be careful here, because the easy move is to say "AI is CurrentC" and collect the applause. I don't think that. Something real is happening — I've written about the job numbers, and I don't think they're noise.
But I'd ask three questions before accepting any confident claim about where this goes.
**Who benefits, and who has to adopt?** Where those are the same person, expect it to stick. Where they diverge, expect friction that the projections don't include, and expect the people making the projections to describe that friction as resistance to change.
**Is this shipping, or is this being announced?** Pilots are not deployment. Partnerships are not revenue. I spent years failing to make that distinction and it cost me most of my accuracy.
**Who is telling me, and what do they hold?** In 2014 the loudest voices about payments were the ones with balance sheets riding on the outcome. That is at least as true now, in both directions — the people forecasting total transformation and the people forecasting collapse are, in the main, positioned.
The one thing I'd bet on
Not that AI fails. Not that it transforms everything.
That the technology question resolves faster than the distributional one, and that the distributional one is the whole thing.
Mobile payments got sorted out inside a decade. The card networks kept the interchange. That was never a technical question, and no amount of engineering was going to change it, because it was a question about who had leverage.
The AI conversation is currently being conducted almost entirely in the language of capability — what it can do, how fast, how soon. Capability will be settled. Who absorbs the cost of the transition, and who keeps the gains, will not be settled by capability at all.
I watched an industry spend five years and an enormous amount of money arguing about technology when the argument was actually about money. I'd rather not watch it twice without saying so.
J.W. Gant
