Last week, Bill Vorhies wrote a provocatively titled first in a series of upcoming blogs on the subject. “Data Scientists Automated and Unemployed by 2025?” Though I’ve considered what impact automation might have on the profession, I’d never considered the question of whether or not there will be a need for data scientists at all.
Recently, I’ve been thinking about the tragic story of John Henry as it relates to data science. To me it is a solid (if extreme) analogy of what I’ve observed in the industry.
I was 8 when I learned the legend of John Henry. My teacher showed a video that included the ballad which has somehow stayed with me through the years (I’ll die with a hammer in my hand). The steam engine seemed so evil. I don’t think I teared up at the part when he died, leaving behind a widow and a young son. Well, maybe a little. Damn you, steam engine!
Last year, I discovered my daughters didn’t know the basics of American folklore so I found a video online and watched it with them. While I still admire the determination and the sheer force of will demonstrated by John Henry to beat a machine, his death is more tragic to me because the race seems pointless. In fact, I’d root for him to lose the contest.
The tunnel where John beat the machine and died may be Big Bend Mountain in West Virginia. It’s at least a mile thick. As the legend goes, John Henry beats the machine and as a result no machine is used to create the tunnel. Instead it is built with the physical labor (and sacrifice) of a thousand men. That sacrifice includes John Henry’s life and the deaths of hundreds of workers who are buried in unmarked graves at the tunnel entrance.
Back to Bill’s article:
During our three-month beta release, we got an incredibly positive response from the data scientists who used the platform. “I can get done in hours what it can take weeks to do.” “Accuracy is great. Sometimes it even beats my models.”
Score! Now we launch, figure out pricing and convert advocacy into buyers.
No surprises on the typical sales objections. However, a couple of times I’ve come across something that did catch me off guard. Generically this is the conversation:
“It’s cool stuff. But I’d rather do this work myself.”
“Really? All of it? You can use it to reduce the mountain of workload and use your powerful brain to work on the super tough/super interesting challenges.”
“Yes. All of it. I just like doing it. I don’t care if it takes longer.”
As a business guy, I’m perplexed. So it’s faster, as accurate, immediately deployable and no thanks? The thought of all the lost productivity is appalling. On top of it all, there is a shortage of data scientist talent. How hard do these guys want to work? 100+ hours a week? Talk about dying from exhaustion!
I’m guessing it’s a matter of pride and fear. My parallels:
Pride: John Henry was the strongest steel driving man ever. He decided “No machine is ever going to beat me.” Data scientists have the most analytical predictive brains ever. In some encounters with them, I’ve essentially heard, “No automated machine learning is going to beat me.”
Fear: John Henry was worried the steam engine might take jobs. The specter of automation might create the same worry in data scientists.
Pride, fear (and laziness) can result in holding on too tightly to the tools of our professions. For John Henry, it’s his hammer. For data scientists it can be using “manual” tools. And for me it’s…
This is a gut check for me personally. Am I unconsciously letting professional pride keep me from adopting innovation? Do I hold onto things that I shouldn’t, just because I enjoy doing them? How about my team? Am I acquiring the tools that best enable them? Am I pushing them to adapt?
An inspirational lesson from John Henry, “A man ain’t nothing but a man. He has just got to do his best.”
In the spirit of being the best at what I do, for making my team the best they can be, I’m embarking on an “innovation audit” first for myself followed shortly by one with my team. We will let go of practices and tools that may be holding us back (even if we are fond of them). We will look for innovations that can make us better: drive product development, optimize marketing mix, identify the best partnerships.
I’m betting my team and I discover a lot.
Will data scientists be unemployed by 2025? As data scientists are currently defined, the answer is yes. And I believe it will happen even sooner. But no profession is exempt from that. It’s true for all of us.
“Change or Die” challenged Alan Deutschman in his Fast Company article.
“I’ll die with a hammer in my hand” declared John Henry in the song.
That’s a tragedy we can avoid.
Check out our platform: www.purepredictive.com/registration