Until recently, a new graduate had a direct path. Go to college, study hard, get a good job. Ambitious students did this for decades, and it generally worked. It’s baked into our cultural DNA.
But AI is rewriting that script, and if you want to stay ahead, you’ll need to think differently. When fields like computer science, medicine, law, accounting, and business all face massive upheavals, where do you focus? It’s not an easy question.
And it’s not just a question for students. Everyone will find themselves squeezed into a similar position, and probably faster than you think. The pace of AI is hard to wrap your mind around. The top AI models are getting remarkably smarter and cheaper to train and operate, and they are not slowing down. AGI will probably be achieved in the next year or so, if it hasn’t already.
How to avoid picking the wrong things
There are two ways to look at this.
One is to choose a field, follow the usual steps, and adapt as AI transforms it. For a long time, this has worked reasonably well.
The second way is to try and front-run where AI is headed and focus your time on things that will become more valuable in the future. This is, of course, hard to do – like trying to catch a wave before it crests.
So how do you avoid wasting time on the wrong things? By focusing on fundamental skills. There are certain skills that will matter no matter how things play out. These building blocks give you leverage. If your goal is to maximize the surface area for opportunity, then you’re much less likely to find yourself left behind.
Let's explore a few of these further.
1. Take agency with what (and how) you are learning
Traditional education will not be able to handle the rate of change, so you will need to work around them to learn faster. Large Language Models (LLMs) are best at coordinating deep learning if you use them correctly. The bar for understanding entire industries is low, even for professionals working in them. Use AI to rapidly build competency in connected domains. I like to use Google’s NotebookLM as a centralized repository for research papers, articles, podcasts and basically everything else I can find on a topic – after enough primary sources are added, I find I can hone in on the key concepts and build a systematic approach to learning. These tools will only improve with time. Use them.
2. Develop the skills required to coordinate systems
AI will automate routine tasks, but critical thinking and complex problem-solving remain highly valuable. Focus on learning skills like critical thinking, problem-solving, data literacy, and creative collaboration. Develop soft skills (communication, leadership, empathy), which are harder for AI to replicate but crucial for managing AI systems.
Think about the architect in a small firm that adopts AI to generate building layouts based on zoning rules and design preferences. While the AI can provide quick drafts, it can’t capture the full context of client needs—such as how a space should feel. The architect’s role shifts from spending hours on drafts to guiding the design process and working within real world people constraints. AI enables the process instead of replacing it.
The difference between AI and past technological advancements is that it doesn’t just make work more efficient—it also provides real-time feedback and guidance, almost like an editorial assistant. AI can help you refine your ideas and decisions as you work. In a world where humans, AI agents, and tools need to collaborate, the key skill to develop will be learning how to coordinate and harness these resources in the right ways.
3. Master the ability to communicate with AI
For most tasks, AI won’t replace humans but will augment them. Once you’ve used AI tools for a while, you’ll notice their strengths—pattern recognition, data synthesis—and their weaknesses, like handling nuance or creativity. The key is learning how to push AI beyond surface-level outputs by giving it better instructions. This means mastering prompt engineering, building workflows with AI agents, and automating routine tasks.
Take writing, for example. AI can quickly generate article drafts or summarize research, but left unchecked, the result feels bland or repetitive. Instead, guide it by crafting prompts that nudge it toward your specific style, then continually refine the output. Over time, you will find more success by switching to more of an AI editor, leveraging the tool’s speed while ensuring the final product still carries nuance and originality. The better you communicate these intentions with AI, the more useful it becomes.
Instead of starting with a written prompt, I have found that by using ChatGPT’s Advanced Voice Mode I get better results by just talking to it for 20 minutes or so on a topic to “start” a draft topic. I then upload the entire train of thought and ask for feedback. This creates a dialogue and a refinement process that is very different than pure input/output. The results are strikingly different.
4. Practice and build creative things
AI can remix existing ideas but has trouble with true creativity. That’s your edge. To create something new, you need to connect and build ideas in original ways. Learning by doing is is like a super power—it forces you to grapple with edge cases and develop a deeper understanding.
LLMs are, at their core, prediction engines. They generate outputs based on patterns from prior data, but genuine creativity requires combining and remixing ideas in ways that machines can’t predict. To make interesting things, you need to build your own internal library of creative references through experience and continual exploration and use AI to bring them to life.
While this has always been important, AI tools now make it easier than ever to get started in creative fields. The catch? With creativity more accessible to everyone, the goal posts for what is considered good will shift along with it. There will always be a market for good taste.
5. Start with first principles
A common mistake is to try and retrofit AI onto existing systems. But problems can now be solved in ways that were impossible before. You’ll never see those opportunities if you’re stuck thinking about how to optimize what’s already there. Most people see AI as a tool to improve existing processes, like adding a faster engine to an old car. Instead, ask: If AI were at the core, how would I design this differently?
This isn’t easy. Humans are naturally anchored to what exists. But if you were starting from scratch today, would you still design schools, hospitals, or companies the same way?Probably not.
6. Think of AI as your co-founder
In the past, if you wanted to build something big, you needed infrastructure—offices, servers, employees to handle everything from marketing to operations. Now, with AI tools and agents, you can automate large parts of that work. You can launch and run a business that would have required a whole team just a few years ago.
AI is making the “one-person company” a reality. This isn’t just about efficiency; it’s about leverage. You can build faster and iterate quickly because you’re not weighed down by the overhead of traditional scaling. A small team can punch far above its weight, using AI to handle tasks like customer service, product development, and data analysis.This inverts a lot of assumptions in traditional competitive environments, but particularly in regards to traditional pricing and business models. Pay attention to any industry that is billable by the hour today. The Red Queen effect is real.
Onward is the only direction worth traveling
So where does that leave us? The best way to prepare for a world shaped by AI is to stay adaptable, keep learning, and build things. AI is moving fast, and your biggest advantage is curiosity and the ability to adapt.
There’s no step-by-step guide—just experiment, take risks, and stay open to new ideas. High agency matters more than ever. Technology shifts start with individuals, not institutions. You have the advantage of being able to move quickly and carve out your own path. Progress is less about playing by the rules and more about figuring things out as they unfold. After all, there are no cheap tickets to mastery.