Enjoying not being "the sexiest job" any more
FINALLY
Happy Thanksgiving in the US! Thursday posts are for supporters of the newsletters, with more off-the-cuff thoughts from the past few weeks.
At this point, in late 2023, it’s been a bit of an worn industry meme for a while now that in the 2010s, “data science is the sexiest job”. The hype train has firmly shifted over to the AI space away from “just data”. This holiday break, I’m willing to bet money that NONE of you are going to avoid having at least one conversation involving “AI”.
Right now the world is in a transition period where some people want to have the AI/ML branding right inside the title like in “ML Engineer”, while others are sticking with the more generic “data scientist/engineer” series of titles. My guess (and hope) is that this gap will start to widen since “AI stuff” has a big enough distinct body of knowledge around it to justify splitting off into its own thing compared to many of the other data scientist roles out there. Job titles around data have always been a hot mess, with competing tiers of roles that use “analyst”, “scientist” and “engineer almost at random, even though the job descriptions can be extremely similar. Specialization is desperately needed.
As someone who has always firmly been inside the product analytics, user research, data engineering part of what’s now considered “data science”, I can’t wait for the AI jobs to complete forking off into their own domain over the course of 5-15 years. It was always really weird, and rather irksome, to have a significant portion of people entering data science openly declaring they want to build ML models and don’t care about “the other stuff”. The roots of data science were literally founded on the fact that they’re people who understand BOTH the technical data stuff as well as the business stuff. Hell, even most software engineering roles require at least some basic understanding and consideration of business and design realities to be successful.
Now those people can hopefully find a happy home doing the model-building they love, while saving me the trouble of having to wade through hundreds of resumes/job descriptions that are looking for someone who does/doesn’t fit that mold. As more companies get on the AI bandwagon and adopts separate titles for that space, it should make every job searcher’s life easier.
Essentially, I’m just celebrating an industry-wide expectation-setting event.
Meanwhile, because the higher education departments move EVEN SLOWER than the job market itself, we’re going to be riding a wave of people coming out of new-at-the-time data science university programs for another decade as kids picking their majors today slowly wind through the education system and enter the job market. I don’t know what they’ve been promised going into those programs, but I hope that when they come out the other end the job market is clearer.
And at the end of the day, for all of us practitioners who’ve already been in the field a couple of years, the wave passing us over means we can live our lives in relative peace. Job title clarity, better job descriptions, school curricula only affect OUR lives once in a long while if at all. Most of us rightly don’t need to care.
But much like how being a “software engineer” cooled down after the first dot-com bust, expectations of people who aren’t in the field will cool. There’s already fewer and fewer breathless articles about how wonderful the job is. People are actually talking about the shitty parts of the experience and even trying to improve things.
There’ll be fewer influencers churning out “How to be a data scientist!” content trying to monetize the hype (I really really hope). There’ll be less “I’ve had this job for 6 months and all I do is dashboards and I hate it” reddit threads. There’ll be more space and air for us to actually figure out what the heck our profession is and what that identity entails.
We get to grow up a bit!
And so, I’m looking forward to being increasingly less sexy as the years go by.