It's true banks don't work like tech companies, it is also true that tech companies don't work like bank and the 100% of them failed to replace banks, because banks are more trustworthy when it's about client'money safety.
legacy systems are a good thing, it means they already solved a lot of edge cases and been working good enough for years (or tens of years)
The new shiny toy might look good, but if it does 1/20 of what's needed by the business, it's useless.
I work now in insurance, evolving legacy systems to new technologies, legacy systems suck, but they are also rock solid and a lot of things they do are vital for the company and, more often than not, save people's jobs.
A lot of rules we don't understand are there to avoid abuses
I'm working on a system that's offline from 11pm to 7am to avoid that agents created insurance policies over night making unfair competition to real agencies with offices, employees and regulated opening hours.
It's been put there probably 20 years ago, it looks silly now, but it served a purpose
In one year I've been able to replace alone maybe 1/30 of what the main legacy system does. Which, btw, sometimes still calls an even more legacy system on the backend, just like my system exposes what the new system still doesn't handle as a safer wrapper to the legacy one.
It's a good pace, we have proved it can be done without breaking the workflow, now it's time to accelerate and build a little team around the project.
Evolution is layered, disruption is a jouvenile sin and usually it is bad for company's workers more than for companies.
I agree there are things both could learn from each other.
I do think established banks have a lot of baggage and entirely the wrong emphasis for a world which runs on software. Some of their old rules still encode useful information, many don't, and it will be hard to disentangle the essential from the moribund.
It's true banks don't work like tech companies, it is also true that tech companies don't work like bank and the 100% of them failed to replace banks, because banks are more trustworthy when it's about client'money safety.
legacy systems are a good thing, it means they already solved a lot of edge cases and been working good enough for years (or tens of years)
The new shiny toy might look good, but if it does 1/20 of what's needed by the business, it's useless.
I work now in insurance, evolving legacy systems to new technologies, legacy systems suck, but they are also rock solid and a lot of things they do are vital for the company and, more often than not, save people's jobs.
A lot of rules we don't understand are there to avoid abuses
I'm working on a system that's offline from 11pm to 7am to avoid that agents created insurance policies over night making unfair competition to real agencies with offices, employees and regulated opening hours.
It's been put there probably 20 years ago, it looks silly now, but it served a purpose
In one year I've been able to replace alone maybe 1/30 of what the main legacy system does. Which, btw, sometimes still calls an even more legacy system on the backend, just like my system exposes what the new system still doesn't handle as a safer wrapper to the legacy one. It's a good pace, we have proved it can be done without breaking the workflow, now it's time to accelerate and build a little team around the project.
Evolution is layered, disruption is a jouvenile sin and usually it is bad for company's workers more than for companies.