Less coding means AI augmentation is the future.
What Jensen Huang's statement means for data scientists.
Hi there, it’s Matt and Juan! 👋🏽 Thanks for subscribing to the Pragmatic Data Scientist! Check out Datalife360 for more thoughts on AI product and strategy.
Read time: 5 minutes
NVIDIA CEO Jensen Huang recently shook up the tech world by declaring that coding skills are less crucial for our kids' future than domain knowledge.
This is a major shift in the tech landscape if we are paying attention.
Kids means the future. They will be the next data scientists, data engineers, and the people you will lead. There will be a transition. If you are a data scientist now, you will be taking part in this. It's time to start thinking about the future.
You're probably asking yourself, how you will lead the new generation of data scientists? If coding isn't the main focus, what is?
In this article I will answer:
Why is augmentation the role of future DS?
What does this mean for the data scientist role?
The role of the data scientist will evolve significantly in 2024 and beyond, diverging from the "sexiest job of the 21st century" title Harvard Business Review bestowed in 2012.
We must seize emerging opportunities. We need to equip the new generation of data scientists with crucial skills. Data scientists who learn to do both gain a competitive edge.
Understanding AI's role in augmentation is crucial. We can focus more on work on that harnesses our knowledge as fuel to solve novel problems.
Why is augmentation the future of DS roles?
When Jensen declares that coding skills are less crucial, he's referring indirectly to AI augmentation. It will evolve as AI is adopted.
Barriers to coding are getting lower. As new technologies develop, tasks that used to require a lot of effort, study, and time will become much easier. That doesn’t mean replacement.
For data scientists, it means we need to really think about the value. Coding’s value lies as a tool. It’s a mean to the end. As John Carmack stated, our core skill is problem-solving. Don't lose sight of the essentials. The real question should be: can you "immediately adopt and leverage" these technologies?
Augmentation isn’t replacement for hard data scientist skills.
Without fundamentals its a cart without a horse. You certainly can still pull the cart without the horse. But if you have a horse, you can pull faster and carry a bigger load - with less effort.
You’ll still need to know computer science skills. You’ll still need to think about statistics. And yes, you’ll still need to build, plan, and think. But augmentation helps you build efficiently - if you have been working on the fundamentals.
Every business has its own mundane and repetitive work to handle. Even before generative AI and copilots, ML was used to enhance existing manual tasks. AI augmentation will take it even further.
Immediate adoption of AI augmentation is already happening. In the legal field, Macfarlanes uses AI to support research, analyze documents, and create drafts, with human lawyers reviewing its work.
Lawyers aren’t disappearing because of AI. And it’s no replacement for their legal skills. But AI augmentation enhances the mundane legal processes and saves them time that’s valuable elsewhere.
We data scientists are no exception.
Copilots will play a major role in the transformation of the DS role. There are several data scientists I know who are already using augmentation to improve their work. From creating docker files to implementing a framework for an API, from streamlining feature engineering to refining recommendation algorithms to enhance user engagement.
And this is just the beginning. As industries adopt AI, they will face proportionally complex problems. That adds entropy to the development process. There are more dependencies, technical challenges, and product needs.
After AI augmentation helps us manage this complexity? We can take on more impactful positions, and become business partners, architects, and strategists.
But before that, we need to know how to integrate into our work.
What does this mean for data scientists jobs?
As AI augmentation diminishes the coding burden, we're at a pivotal juncture.
We need to bridge out technical skills with a strong sense of business strategy. It's also crucial to make the most of easier access to our field and to take serious responsibility in teaching those who come after us.
These efforts are key to shaping the future of data science as AI grows more influential.
We need to focus bridging the technical and business gap.
As AI grows, so do the challenges we face. To keep up, deepening our domain knowledge becomes crucial.
This means adopting a mindset focused on first principles, asking ourselves:
Why are we solving the problem?
What must we build?
How do we plan to build this solution?
Domain knowledge isn’t just sitting in a position. Its dynamic. Domain knowledge evolves alongside tech. Its actively trying to understand the domain.
The top data scientists and engineers don't just work with data—they live and breathe the questions that drive their projects forward. This relentless curiosity hones their expertise. They become leaders in crafting AI technical solutions, products, and strategy by asking these questions.
Domain knowledge evolves alongside tech. We need to evolve alongside it to bridge the gap.
AI augmentation won’t answer these questions. We do.
Getting to the heart of these questions doesn't just improve our workflows. It transforms them. Augmentation with our knowledge and experience helps speed it up. This leads to quicker projects and more efficient processes. Over time, we build a pace of work that gives us an edge over others.
By focusing on these questions, we connect business needs with technical solutions. By using augmentation, we can focus on faster quality delivery.
We need to ensure AI augmentation helps us solve business challenges. It should make us quicker and more effective.
We need to navigate Lower Barriers to Entry
Augmentation can be intimidating - it lowers barriers to entry. Throughout history, devices like the cotton gin and the personal computer have done this. Generative AI is no exception.
Less experienced workers can work faster with AI augmentation. They can work through tasks that took us hours. With enough experience and domain knowledge, they may even upskill faster than we did.
Great AI augmentation extends your skills, knowledge, and experience.
This doesn’t mean we should stop learning or be discouraged. But we need to account for augmentation as a factor. This is an opportunity if we approach it right.
Think of augmentation as an extension of you. We should not see AI augmentation, such as copilots, as replacements for data scientists. We do need to see them as extensions of our current skills and roles. Augmentation is a natural evolution of tools and process. It enhances our ability to solve problems and find insights. We use a cart to move boxes and furniture. Augmentation works similarly.
Redefine your expertise. We need to rethink domain expertise. The game is changing. It's less about technical tasks and more about solving tough problems. Now, your value needs to also center around crafting solutions to complex ML problems. As juniors take on routine work with AI's help, your role evolves to leading with wisdom. You set the standards, mentor, and uplift the team. That can’t be automated. Only augmented and enhanced.
Use augmentation as lever. Augmentation is a game changer. It automates the dull, menial and repetitive tasks, and may even standardize coding if done right. This frees up time. It means existing team members and new hires are onboarded faster. Augmentation frees up time for more training and greater project responsibilities for them. It's a transformative shift in how teams come together and work.
Lower barriers to entry don’t diminish your role as a data scientist. Approaching tech as means, rather than an end can make your team more productive.
The great future staff and lead data scientists will figure out processes for augmentation to push their careers farther.
If we want to use AI augmentation to push ourselves and our teams further? We need to make sure our teams are in the right place. We need to mentor the next generation.
We need to mentor to leverage AI augmentation.
At a certain point in your data scientist career, you come to a realization:
“I can’t do this all on my own. I need help.”
Individual skills are important. But so much of it means you need to have a solid team behind you. Teams that have well defined AI augmentation workflows and mentorship will be far more productive than those that don’t.
Code is just a shovel used to create a cathedral that is a valuable AI solution. You can leverage augmentation to reduce junior's workloads - and focus them on building quality solutions to the problem.
You need to mentor and upskill your team. Augmentation can make you productive. But you need to help your team members to be productive too. You need to mentor AND leverage AI augmentation to upskill them.
Your experience is valuable here. Mentoring builds frameworks for your team. Teach them to use augmentation the right way to learn more, build faster, and create solutions. Having a good delivery cadence helps morale and team cohesion.
Learn to provide best practices around AI augmentation to let them do it the right way. Yes, generative AI will generate a lot of less-than-ideal coding habits in our reports. That’s something I’m seeing more and more. But you can use the time freed up from writing yet another YAML or Dockerfile to mentor them to break it.
Returning time is a big benefit from augmentation. It gives lead data scientists more time to mentor. It gives their reports time to be trained and ask questions. Augmentation is great deal if lead data scientists leverage it for training and mentorship.
Teams led by these leaders will stand out - they’ll be far more productive and deliver faster. Complex business problems require cohesive teams, who know their role and can leverage AI augmentation.
Make augmentation work for your team. Use mentorship to enhance it. Use both to build great projects and add value to your reports work.
Final Thoughts
Jensen Huang isn’t saying the need to learn computer science or coding is over. He’s expressing that adopting and integrating AI augmentation is coming. While he implies that there’s no need to pursue a computer science degree, he doesn’t say that knowledge and fundamentals don’t matter.
Data science teams and their leads need both to succeed. AI augmentation gives them time to upskill, create better solutions, and use domain knowledge to guide both.
Domain knowledge and problem solving will be the core going forward. Especially for data scientists in the generative AI era. AI augmentation can free up time to solve problems and develop our teams.
With generative AI, the barriers to programming and technology have been significantly lowered, making it possible for almost anyone to become a tech expert.
AI has changed the way we interact with technology, bridging the previous technological divide.
With barriers lower, we need to think how to use it to augment your workflows, existing knowledge, and strengthen your fundamentals.
We need to think about:
What unsolvable business needs are possible with AI augmentation?
What AI and data products will you enable?
How can you reduce cost and increase productivity?
Why do you need it?
Every new tech wave creates opportunities. Find those and develop the skills to make the most of those opportunities.
AI augmentation is a force multiplier. But it can only enhance the skills fundamentals, and experiences already there.
 | A guest post by
|
 | A guest post by
|