You don't need to solve hard, technical problems: From an Oxford University Engineer

July 15, 2025 9-minute read

When I started as a student, I wanted to solve “hard technical problems”. After all, it’s what I was good at in school and what interested me. The process of learning new maths or physics models and playing with them in my head to understand their consequences excited me - often a 3blue1brown video and a concept I didn’t understand were all I needed. However, once I started working I realised that most problems that warrant a solution are not technical. Furthermore, I saw how learning the shiny new tool that humanity just came up with, such as machine learning or AI will not save the world. Actually, a lot of key problems are solved with boring technology, so that you can focus on the several other dimensions these problems have. This took me more than a year to fully understand after reading Ben Kuhn’s original article, so by explaining my understanding of problems beyond simply the “hard and technical”, hopefully you can understand this message too.

The issue with solving problems after school: Not every problem has a solution

For most of your education, most of the hard problems you’ve solved have a solution or even a score. This makes sense because the scores from tests are used as a measure of your ability compared to other people. Even the hardest problems you might have in school, such as maths competitions and science olympiads have clear solutions. Even the projects I had in Oxford had a rubric, like 15% of the grade for presentation, 15% for introduction and conclusion and 70% for engineering content.¹

The issue is that real world problems are nothing like this. Even defining your problem can be a struggle. Defining how to measure the quality of your solution is another struggle. In fact, during my senior thesis I spent more than two weeks figuring out the best way of measuring the quality of an AI generated sentence for predicting natural conversation. This is common theme that when solving an interesting problem with a technical solution we have to provide a proxy problem instead. For example, language models are typically trained to predict the next word in a sentence, rather than generating language in a broad sense.

The issues with technical problems

Human problems matter to people, not technical ones

All the problems that matter to people are human problems, we rarely care if the solution is technical or not. This might appear to conflict with what you think is an intellectually interesting problem. This conflict is more common the closer you get to pure maths or abstract theory too, where applications become less clear. Additionally, your incentives might include trying to find completely novel problems or unexplored territory, believing this is how you make your mark.

However, instead of solving a problem somebody has never thought of before, if you solve a problem everyone has thought about before you will achieve more impact.² The biggest human problems are universal and everlasting. Furthermore, the more universal the problem, the more impactful the solution. At an extreme, consider the problem “what is the meaning of life?”, religion often provides an answer and it has given many people value.³

It pays to have an interest in understanding what problems people actually face. One framework to get you started is health, wealth and relationships. I encourage you to think of your own too. You will start seeing problems everywhere-from “how do we hire the best person?” to “how do I have a good holiday” and “how do I find a romantic partner?”. Many successful companies are built on providing better solutions to these eternal challenges in our ever-changing world.⁴

Non-technical skills provide leverage for big problems

Say that you find a problem you want to work on. A one person team can only scale so far because you are limited by the amount of hours you have in a day. Hence, to scale impact you need leverage in the form of people.⁵ This involves hiring effectively, delegating responsibility and coordinating collective effort. This is true for any big problem in any industry.

If you’re starting a company, you must assemble a strong team around a clear mission, develop a business model that can sustain growth, secure funding through revenue or investment, and build an organization designed to scale. If you want to become a professor, you must rally teams around research questions, secure funding through grant proposals and coordinate complex projects with multiple contributors. In all paths, leadership, management, and strategic thinking are key.

No great human problem was ever solved individually.

How to start solving real problems: You can make a big impact with boring technology

Given all this information about the problems that affect people how do we go about actually solving these problems?

First, use the best tool for the job instead of the most shiny one. This often means embracing boring technology - reliable, well-understood tools that will free up your energy for solving the problems that are most useful to solve - technical and otherwise. Technology alone is not a silver bullet: one app won’t solve healthcare or poverty. Real solutions require understanding systems, changing behaviors and working within constraints rather than around them.

For most hardworking and intelligent people, this approach will be very difficult. I’ve prided myself on my ability to solve any problem with enough time and effort – but this is not the way to think. You need to choose your battles.

Next, to get started and improve on solving real problems, you should learn how to work hard and how to develop agency and taste. I think one of the most important things is to take the initiative and define the problem for yourself and try it solve it your way. Once you start, make sure to keep moving and think real hard to solve the sub problems you encounter.

Solving real problems doesn’t require the most modern technology

The amazing thing about the present is that technology allows us to innovate with relatively little upfront investment because we need minimal hardware compared to previous generations.

Consider what’s possible: 55 years ago, Douglas Engelbart delivered the Mother of All Demos in 1968, introducing the world to revolutionary concepts that still define computing today:

Video conferencing (connecting people across distances)
The computer mouse (intuitive computer navigation)
Word Processing (more efficient writing)

We live in a fast paced world, so it can be tempting to want a solution as quick as possible. Remember there are no shortcuts: find a real problem and try your best to solve it. When you are in doubt, think about Douglas Engelbart and his team and imagine what you can do with a few dedicated people, some capital, and focused time investment.

Summary

Instead of solving hard, technical problems, focus on solving human problems that people actually face in their daily lives. By using boring technolgy you can create big impact by giving yourself the resources to solve the problems that really matter, which are usually just as complicated and hard.

Additional: Strategising around complex problem solving

Learning how to manage complex problems within a scope is an incredibly valuable skill that could fill books. In business, this type of problem solving often appears as “strategy”, and strategy consulting essentially provides frameworks for tackling fundamental questions such as: “How do I make my business more profitable?” or “How do I make my business more resilient?”

The key insight for technical people is that the solution space for complex problems is enormous. Using a purely rationalist, bottom-up approach to analyze every possibility is not just impractical—it’s impossible. Instead, you need a combination of top-down vision and bottom-up execution. Figuring out ways to navigate such a high dimensional space can be a fun challenge. Here are two thoughts to help:

Complex Problems Can’t Be Solved, Only Navigated

Arthur Brooks said that complex problems cannot be solved, only experienced. His analogy is a football game: you might have a set play for the first 10-20 seconds, but after that, with so many actors and possibilities in motion, you can’t rely on deliberate analysis (System 1 thinking). You must shift to intuitive, adaptive responses (System 2 thinking).

This mirrors how successful leaders approach business strategy. They start with a clear initial plan but remain ready to adapt as circumstances evolve. The goal isn’t to predict every scenario but to build the capability to respond effectively to whatever emerges using your wisdom and experience.

Truman’s Law: Marginal choices compounds

Given the enormous possibility space in complex problem solving, historian Ian Morris introduced Truman’s Law. When asked what makes a great president, Truman reportedly replied: “The truly great president is one who is right 51% of the time.”

This highlights a deep insight about complex problem solving: Having a slight edge in good decisions over time can compound in complex, unpredictable systems.