by Serguey Shinder
Around 2011 I ended up responsible for a batch transfer and reconciliation layer that sat between a broker and a settlement system. Files in, files out, a fixed-width format defined in a document older than my career, and a scheduling mechanism I will not name because naming it would date me precisely.
Nobody wanted it. The interesting work was two floors up on the new platform, and that is where my friends went. I stayed partly because I had been asked to and partly out of a lack of nerve, and for about four years I was fairly sure I had made a career mistake. There was nothing about it I could put on a CV that would mean anything to a stranger.
Then something happened slowly enough that I did not notice it happening. The two people who had built it retired. A third moved to another country. Meanwhile the system did not go away, because actual money moved through it every night and the cost of replacing it was enormous and the cost of leaving it alone was nearly zero.
By about 2019 I was one of a very small number of people who understood it, and I was being asked, politely and with a budget attached, to explain it to people who needed it to keep working.
I want to be careful about the lesson, because there is a flattering version of this that I do not believe. The flattering version is that loyalty to old technology pays. It does not, by itself. What paid was an arithmetic thing. The number of people who understood that system was falling faster than the need for it was falling. That gap is where the value sat, and it had nothing to do with the technology being good.
Which means the same arithmetic can go the other way, and I have watched it do so to people I know. A skill that is scarce because nobody wants to learn it is often also scarce because it is ending. If the need falls faster than the supply of people, you are not sitting on a premium, you are sitting in a room that is getting smaller, and the last person in the room does not get paid the most. They get paid for a while and then not at all.
So the thing I actually watch now is not how fashionable a technology is, in either direction. I do not chase the new and I do not romanticise the old. I try to form an honest view of two curves: how fast the need for this is shrinking, and how fast the number of people who can do it is shrinking.
Being early is comfortable. Being the last one is not, and from the inside they feel exactly the same.
– Serguey Asael Shinder
Leave a Reply