John Carmack Goes On Coding Retreat With OpenBSD

[coder]

I find it extremely alarming that the way the average person's money is handled via credit cards is a trade secret (at least that is what I've gathered from some of your comments).

Of course, this is legacy from the era of security by obscurity.

Do you know of efforts to improve that? Something something blockchain?

Repeat after me: blockchain does not scale (like that). Blockchain is fine for high-value, low-volume transactions, where there's value in having the provenance and full ownership history (e.g. fine art, real estate, etc.), but it's far too inefficient and scales much too poorly to replace all credit card transactions.

I'm not saying we'll never use crytocurrency for everyday transactions, but it's not simply going to be blockchain-based. Maybe the right tech is already out there. I don't closely follow cryptocurrency developments.

[fuzz]

I can't actually get too much into this. I actually tried to do a bit of research online, because if some information is available publicly, I'm allowed to talk about it. The "best" source I've found is https://www.creditcards.com/credit-c...sProcessed.pdf

I find it extremely alarming that the way the average person's money is handled via credit cards is a trade secret (at least that is what I've gathered from some of your comments). Do you know of efforts to improve that? Something something blockchain?

[grok]

The banking world is really not in a "business as usual" phase. Sometimes it collapses spectacularly (see securitization, circa 2008-2009 lol).

What if you could create your own banks..
Open two new banks, one credited with 10,000 EUR or dollars. Have Bank 2 borrow 120,000 dollars from Bank 1. As a private ordinary person, borrow 1,440,000 dollars from Bank 2.
Declare your banks bankrupt, ruined, out of business and run away with the money.

[AndyChow]

The interface between banks and the credit card companies should be a fairly simple API. Otherwise, this stuff just couldn't work as well as it pretty much always has, and the credit card companies wouldn't have been able to entrench themselves in such hegemonic positions.

That's exactly what I'm saying. It is otherwise, and this stuff does work as well as it pretty much always has. THAT'S the miracle.

[AndyChow]

I'm trying to be open-minded. Thanks for taking the time to explain this to us.


We can only hope...

I can't actually get too much into this. I actually tried to do a bit of research online, because if some information is available publicly, I'm allowed to talk about it. The "best" source I've found is https://www.creditcards.com/credit-c...sProcessed.pdf

As you can see there, it details 3 activities, authorization, batching, and clearing. Regardless of what the diagram of step 3 of authorization shows, the text is accurate: "The acquirer sends a request to the issuer to authorize the transaction." The acquirer can be the issuer, for example if you are using a Chase card, and the merchant has a Chase terminal. Or it can be many combinations. But the card network (Visa, MasterCard, etc), floating in the two first activities, isn't actually talked about it the text. Perhaps because they don't need to be involved directly (wink wink).

The reality is a lot more complicated and interesting than what is shown here, unfortunately I can't get deeper into this because I realized a lot of this stuff is 'internal use only' type of info.

But consider that the system, overall, works, and it works really well. You can travel the world and pay for things just about anywhere with a credit-card, and it works. No single 'large corporation' could ever handle it all. There are many payment facilitators and new fintechs are constantly emerging, challenging the structure and trying to provide a data edge or improve capital fluidity. The quantity of data handled is unbelievable, and the real fun starts you start thinking of how to efficiently use that data to drive management decisions.

The banking world is really not in a "business as usual" phase. Sometimes it collapses spectacularly (see securitization, circa 2008-2009 lol).

[coder]

What I am saying, is that the credit card companies have actually very little to do with handling the transactions. I can't get too technical, because I'm a bit dancing between what is public and what is proprietary knowledge. But in actual transactions, credit card companies have little to nothing to do with the actual transfers.

I'm trying to be open-minded. Thanks for taking the time to explain this to us.

If tomorrow, both Visa and Mastercards infrastructure were to temporarily crash, it wouldn't block any transactions, and the effect would be minimal.

We can only hope...

[drSeehas]

... Most banks still run Fortran code from the 80's in their mainframes, ...

???
You mean COBOL, not FORTRAN?

[AndyChow]

What you paint as the complex part (i.e. the interworking of all these systems) is the part you dismissively ascribe to the credit card companies. I think that's what Starship & I are saying.

The interface between banks and the credit card companies should be a fairly simple API. Otherwise, this stuff just couldn't work as well as it pretty much always has, and the credit card companies wouldn't have been able to entrench themselves in such hegemonic positions.

What I am saying, is that the credit card companies have actually very little to do with handling the transactions. I can't get too technical, because I'm a bit dancing between what is public and what is proprietary knowledge. But in actual transactions, credit card companies have little to nothing to do with the actual transfers.

As to why the credit card corporations have been able to entrench themselves... I'll offer a guess. It's only a guess, I really don't know the real answer. My guess is they offer a neutral bridge. And they also collect all data, so can provide non-biased aggregate market share information.

Let me put in to you this way: If tomorrow, both Visa and Mastercards infrastructure were to temporarily crash, it wouldn't block any transactions, and the effect would be minimal. The topology and infrastructure is much more resilient and independent of them than what you've been let to believe.

And the market is evolving. Yet the old backbone was built well, and can handle it. Old code, built in the 70-80's, that scales and only grows more powerful with new hardware. Compare that with new software that ages poorly. That was my initial remark.

[coder]

Speaking of fast cars, Bill Gates actually got his license suspended, in the early 1980's, for speeding in his Porsche.

Carmack famously had a bit of a thing for exotic cars, as well.

[coder]

Humm, no. I can't blame you for not knowing how the field works, but terminals are owned by banks, of the acquisition supplier. This is not credit cards. There are the issuers (banks), which push the money, the acquirers (banks mostly, a few other players), and intermidiaries (credit card companies, interact (people do use debit cards), and stuff like ApplePay and AndroidPay). When your credit card transaction goes through, it goes throught at least two parties, and three systems (say minimally, bank, credit card, then bank).

To make it more complicated, often credit card companies offer both branding (the credit card you have in your wallet may have the brand Visa, MasterCard, but it also has the name of your bank or a company where you applied for the credit card (ex: Amazon's Chase card)). So it gets a lot complicated very quickly.

What you paint as the complex part (i.e. the interworking of all these systems) is the part you dismissively ascribe to the credit card companies. I think that's what Starship & I are saying.

The interface between banks and the credit card companies should be a fairly simple API. Otherwise, this stuff just couldn't work as well as it pretty much always has, and the credit card companies wouldn't have been able to entrench themselves in such hegemonic positions.