Research is not just about ideas

I used to think research, especially in my fields of interest, was about finding interesting ideas and building logical arguments around them. If a model was internally sound, mathematically motivated, and pointed toward new questions, that felt like progress.

That view is incomplete, and the people who showed me so were not physicists.

Over the past few years, talking to startup founders, VCs, and people running real systems forced me to confront something I had never fully internalised:

ideas are cheap; execution under constraints is the real work.

This is obvious in entrepreneurship. It is far less obvious in research, especially theoretical research, where progress feels detached from timelines, resources, and implementation.

But once you see how founders think, you cannot unsee it.

From intellectual curiosity to execution discipline

Entrepreneurs live in a world where the question is never simply “Is this interesting?”

The real questions are:

  • Can this actually be built?
  • Can it survive real-world constraints?
  • Can it be pushed far enough to matter?

What surprised me was how directly this maps onto serious research.

A research idea that cannot survive mathematical scrutiny is not different, in principle, from a startup idea that cannot survive market constraints.
Both collapse under reality checks.
Both require iteration.
Both demand resilience.

The founders I met were not necessarily the smartest people in the room.
But they were often the most persistent.

They treated failure not as a verdict, but as information.

That attitude completely changed how I approached research problems.

Risk, uncertainty, and long horizons

One of the biggest internal barriers I faced in research was the fear of spending months, or years, on a direction that might fail.

In entrepreneurship, that risk is not theoretical.
It is assumed.

Founders routinely invest time, reputation, and capital into ideas that may never work.
Yet they proceed anyway, because they operate under a simple principle:

uncertainty is the default state of building anything meaningful.

Once I started viewing research through that lens, something shifted.

A failed derivation stopped feeling like wasted effort.
An abandoned approach became part of the search process.
Exploration began to feel less like gambling and more like engineering.

Research, I realized, is not just discovery.
It is construction.

The builder’s mentality in science

Entrepreneurship instills a very particular mindset:

  • You do not wait for perfect conditions.
  • You move with incomplete information.
  • You refine continuously rather than aiming for flawless first drafts.
  • You accept that the path will change mid-way.

This mindset translates surprisingly well to research.

In theoretical work, we often imagine breakthroughs emerging from singular moments of insight.
In practice, they usually emerge from sustained iteration, failed attempts, and gradual refinement.

The founders I interacted with were masters of iteration.
They built, tested, discarded, rebuilt.

Watching that process made me realize that research should function similarly.

A paper is not a monument.
It is a prototype.

Conviction without rigidity

Another lesson I learned from founders was the balance between conviction and flexibility.

They believed strongly in the direction they were pursuing, otherwise they would never endure the uncertainty, but they were also willing to change tactics, models, or execution strategies when reality truly demanded it.

In research, this translates to something subtle but important:

you must believe a problem is worth solving,
without believing your current approach must be the one that solves it.

That distinction alone prevents intellectual stagnation.

The shift in my own approach

Today, when I look at research problems, I find myself asking different questions than I did earlier:

  • What is the minimal viable formulation of this idea?
  • What assumptions am I implicitly depending on?
  • What would invalidate this approach quickly?
  • How can I iterate faster on the mathematics?

These are not traditional research questions.
They are builder’s questions. They are also the questions an investor asks before writing a cheque: what has to be true, and how quickly can we find out that it is not?

But they make the research process far more grounded, and far more resilient.

Research as venture building

The more I think about it, the more I see serious research and venture building as structurally similar activities.

Both involve:

  • exploring uncertain spaces,
  • constructing models that may fail,
  • committing to long timelines,
  • and operating with incomplete information.

The difference is mostly in the medium (equations instead of products, proofs instead of revenue) but the underlying dynamics are very aligned. The allocation question is identical either way: what is the cheapest experiment that tells me this direction is dead?

Talking to founders did not make me treat research like a startup.
But it did make me treat research like something that must be built, not just imagined.

That shift has shaped how I work more than any single technical insight. Ideas I can generate cheaply. What I optimise for now is the rate at which I can kill the wrong ones.