We are officially living in an age of rapid technical progress, and AI easily doing a code monkey’s job is a very real, and very smart, proposition. For people entering the data field, people looking to upskill in order to stay in the field, or people who don’t want to work for a company steeped in legacy, I’ve recommended three skill sets.

1. Be an expert in tools and platforms. I heard something astute from Gary Vee: “Technology is undefeated.” You need to be able to navigate which tools should be used for what, with what, and for how much. Snowflake or Databricks? How can this IDE connect to git? You also need to know how to read and understand a user agreement, because I have worked with a boss who thought we couldn’t use open source tools to run our reports. It took two of us to explain what the commercial use section meant, which took up time that could’ve been used to do some actual work.

2. The ability to communicate to technical and nontechnical people. I worked in the tech industry for years and I ran into a surprising number of technological dinosaurs. One thing that’s going to happen is that the computer literacy divide will widen - nowadays the extent to tech savviness of a large chunk of people is how to send messages and look at TikTok on their phones. Most people lack the knowledge of how to be resourceful to solve their problems. Yes, you might have to explain statistical results to managing directors who don’t know how to set a calendar reminder. But most likely, you will have to communicate with clients who lack the technical knowledge to describe their problem. You have to parse this information through iterations of questions, hearing what wasn’t said, and testing tangible solutions. To gauge your ability now, practice explaining in two jargon-free sentences what a p-value is, then explain it again in a different way, and lastly, make sure your explanation is aligned with the American Statistical Association’s Statement on Statistical Significance and P-Values. The same boss (a director of statistics) from part one didn’t really know his stuff, so whenever a colleague or client of ours would ask for clarification on a statistical concept, he would purposefully answer with long, jargon-laden responses in order to obfuscate them into responding, “um… ok.” Be the person your teammate, or paying client, wants to ask follow-up questions to. (Pro tip: draw a picture for them.)

3. The ability to correct and supervise AI. Where are the biases? Was the underlying data correctly cleaned? What edge cases are missing? Was the right question asked? Finding out the right problem is a function of mastering your communication skills, so reread part two. If you are a career professional and haven’t been able to answer any of these questions with your own work, I hope you are working for eoraptors less savvy than you, and then stay there until retirement.

A perhaps unexpected way to learn all these skills is to learn design, or at least design thinking, and become good at it. The point of design is to make something that people want to use and like to use. Having a user-centric approach to solving problems will set you apart from people who limit themselves to “what I read in a textbook years ago.” We are in a world where personalization, or at least the illusion of it, is important.

Listen. This might be an unpopular opinion, but deep knowledge in running machine learning models won’t be as widely necessary in the future. AutoML came out years ago, so why isn’t it time for computers to build and test models? Like I said before, you just need to work on iterating over the right question, which ChatGPT can’t do yet.

The last thought: If you don’t know how, then learn how.

I had a good friend who wanted to make income outside of his work at the gym he worked at. We developed a social media plan for his two strengths: being bilingual in Spanish and English, and his ability to create a killer daily workout. He loved what he did for his clients in his gym, so our focus was to scale that service to a global level.

1. Being bilingual - I suggested specifically for him to create content in both languages on the same channel. There’s a lot of English-speakers learning Spanish and Spanish-speakers learning English. If that isn’t what the market wants, then he could easily pivot to whichever audience is attracted to his content.

2. Pre-made routines. I, and friends my age, have no time or interest to come up with a routine everyday. Just tell me what to do and I will do it. This is a great opportunity to connect with busy people who go to the gym and need a resistance program to follow. People like me have a lot of problems to deal with practically every day, but keeping in shape is imperative to keeping my energy up. If someone else can program my daily routine, I will be a fan for life!

The first step was to create a consistent posting schedule, no matter how small. Since Monday was legs day, I suggested that he post a quick “how-to” video on one move every week.

The easiest platform to make money is YouTube, because he could get paid directly from AdSense, whereas on TikTok or Instagram, the grind could be a little bit more difficult to find and maintain promotional partnerships. Of course, “easy” is relative. It would most likely take more than a year of consistent posting to see a hefty paycheck. But doable because the market is there.

Future considerations: like any other strategy, this is an experiment which may lead to unexpected results or shift in priorities. Where I could see this taking him could be developing daily programs for trainers at other gyms, or maybe even having a few online clients where he programs their weekly schedule.

The key is to find your own special thing that people consistently come back to you for. This sweet spot is your ikigai. Keep at it and you will find what makes you feel fulfilled is to make others happy.

I had a client in the manufacturing industry who needed to hire a data person/team to reshape their internal data organization, with the long-run objective of getting his company to the McKinsey Lighthouse Network. His first thought was to poach someone from one of the companies in the directory to help them get to that place. But I pointed out, if you are looking for someone to apply a tried-and-true formula, then you are behind in the game. I then asked, if you hired someone from an entirely different industry, or someone who has hopped around industries, would your senior management listen to his or her out of the box ideas? Or are they too settled in their corporate ideation comfort zone?

My recommendation to him was to focus less on getting someone with a singular background in manufacturing and look for a new hire based on skills only. The main skill for his company’s needs was data architecture with an emphasis on a deep knowledge of tools and platforms. Getting his company to be recognized as an innovator would be a different ball of problems I didn’t think they would be ready for. I found it hard to believe that in their history they never had any team members who could get the organization to the next level.

If your top talent keeps leaving, then your environment (ie, middle and senior management) is not ready for innovation. What happens when someone unexpected suggests something that seems very unlike what your company is doing now? For most places, I think the primary reaction is to be dismissive. So stop trying to sign on superstars with the expectation they will launch you into the stratosphere by playing it safe.