To celebrate the 100th TSQL Tuesday, this month’s topic is what things will be like 100 months from now.
Since we are prognosticating, I want to take a guess at one of the constraints limiting the future. I present you with Meidinger’s law:
An industry’s growth is constrained by how much your junior dev can learn in two years.
Let me explain. On my team, one of our developers’ just left for a different company. We also have a college student who will be going full time in May, upon graduation. How long do you think it’s going to take the new guy to get up to speed?
And how long do you think he’s going to stay?
The first number seems to get longer and longer. The second number seems to get shorter and shorter. What is going to happen when the two numbers meet?
Every two years, give or take
Cameron Keng at Forbes thinks you should change jobs every two years. In my opinion, if you change jobs any more frequently than that then any manager worth her salt isn’t going to hire you.
It certainly feels like a lot of people change jobs every two years. I see a lot of turnover in IT. In reality, the median tenure for computer jobs is has been 4.5 years for the past decade:
Maybe a bit closer to 3 years, when talking about people in their mid twenties.
And what about developers in particular? Well, shit.
We are all imposters
If you are going to change jobs every two years, then that’s how long you have to learn before it’s on to the next thing.
And I think Meidinger’s law is our saving grace in some sort of way. Things keep changing, we have to keep learning. But how much we actually have to learn is constrained by that 22 year old drinking red-bull with no family obligations.
Oh yeah, we were supposed to be predicting the future. Well I think the fact that we are all going to be replaced by that 22 year old some day gives us a hint at the future.
I think more and more things are going to be abstracted away. We’ve seen it with virtualization. We’ve seen it with the cloud. These abstractions mean the new guy can learn new, more important things. He doesn’t have to be intimately familiar with RAID 5. He doesn’t have to have the OSI model memorized.
I think we are seeing it now with data science and machine learning. So much of those areas require a Phd and years and years of study. But things like Azure machine learning and Azure cognitive services are going to get easier and easier. So easy that even the new guy can do it.