I read the piece but IMO this there is a hidden last step not mentioned:
Present in the article:
1) Core banking was good in the 70s - True
2) Core banking sucks now - Also true
Missing:
3) The switching costs for a major bank are insanely high and no regulator (to my knowledge) has given buy-in for an easier regulatory framework while these systems are moved. The result is that you will have to run 2 books while you transition over and work out all the kinks, then also need to port all the historical data. There is no management team that is going to okay this spike in operational complexity without the regulator giving them a hand.
> Here is the most telling evidence that core banking as an architectural philosophy is obsolete: virtually every new bank built in the last decade has explicitly rejected it.
>Monzo, Starling, Revolut, Nubank, Chime, TymeBank, and dozens of other digital banks have built around domain driven design and event driven architecture from the ground up.
Every single one of these neo-banks has so far only done the "easiest" part of banking, i.e offering a checking account, some savings/investments and maybe basic consumer loan. The current regulatory environment heavily favors universal banks and the problems start once you try entering into commercial/IB/mortgages.
I have friends at UBS and just moving all the clients from the Credit Suisse platform has been a massive shitshow in all departments and already went past their internal deadlines.
The exact same architectural pathology exists in enterprise visual systems.
We correctly diagnose 'core banking' as an unmaintainable big ball of mud because it lacks rigid, compile-time constraints. Yet, we somehow tolerate the exact same entropy in how B2B enterprises build their digital interfaces and brand assets.
When a design system relies on post-facto human validation rather than discrete mathematical parsing (Parse, don't validate), you don't get a scalable system—you just accumulate visual technical debt. The root cause of decay, whether in a banking backend or a brand's data topology, isn't age. It's the absence of absolute axioms from day one.
I read the piece but IMO this there is a hidden last step not mentioned:
Present in the article:
1) Core banking was good in the 70s - True
2) Core banking sucks now - Also true
Missing:
3) The switching costs for a major bank are insanely high and no regulator (to my knowledge) has given buy-in for an easier regulatory framework while these systems are moved. The result is that you will have to run 2 books while you transition over and work out all the kinks, then also need to port all the historical data. There is no management team that is going to okay this spike in operational complexity without the regulator giving them a hand.
> Here is the most telling evidence that core banking as an architectural philosophy is obsolete: virtually every new bank built in the last decade has explicitly rejected it.
>Monzo, Starling, Revolut, Nubank, Chime, TymeBank, and dozens of other digital banks have built around domain driven design and event driven architecture from the ground up.
Every single one of these neo-banks has so far only done the "easiest" part of banking, i.e offering a checking account, some savings/investments and maybe basic consumer loan. The current regulatory environment heavily favors universal banks and the problems start once you try entering into commercial/IB/mortgages.
I have friends at UBS and just moving all the clients from the Credit Suisse platform has been a massive shitshow in all departments and already went past their internal deadlines.
The exact same architectural pathology exists in enterprise visual systems.
We correctly diagnose 'core banking' as an unmaintainable big ball of mud because it lacks rigid, compile-time constraints. Yet, we somehow tolerate the exact same entropy in how B2B enterprises build their digital interfaces and brand assets.
When a design system relies on post-facto human validation rather than discrete mathematical parsing (Parse, don't validate), you don't get a scalable system—you just accumulate visual technical debt. The root cause of decay, whether in a banking backend or a brand's data topology, isn't age. It's the absence of absolute axioms from day one.