Researchers, Builders, and What I’m Paying Attention to in 2026

Over the last year, I’ve found myself increasingly focused on a question that sits underneath many others I care about: "can you predict what people will achieve by how they act before any outcomes are clear?"

This lens shows up in how I think about incentives, ownership, intensity, and long-term commitment. I spend a lot of time working with researcher-founders coming directly out of universities – often pre-product and deep in the exploration phase – and I've learned there are signals that present very early. Long before there's been any external validation, you can get a sense of how someone will take ownership, take risk, make decisions under ambiguity, and involve others.

That’s where the often-flattened distinction between researchers and builders becomes useful, not as a hierarchy, but as a way to understand orientation.

‍

Orientation matters more than credentials

The lazy version of the “researchers vs builders” framing suggests that builders ship and researchers think. That misses the point.

The real difference is how people relate to uncertainty and ownership.

Researchers are trained to optimize for truth: correctness, novelty, and contribution to shared knowledge. Builders are trained, sometimes painfully, to optimize for coherence: aligning decisions, incentives, and effort toward something that works in the real world, even when proof is incomplete.

Both are valuable. But they lead to very different behaviors when there is no roadmap, no evaluation function, and no external validation which is exactly the environment of the earliest-stage company.

‍

Why this matters so early

What I’m looking for, especially early, isn’t a “researcher” or a “builder” in the abstract. It’s whether someone with a research background begins to exhibit builder behavior before they’re forced to. This means taking ownership under ambiguity, making decisions without permission, and designing systems for others long before a company exists.

At the pre-product stage, there are no conventional signals to lean on. No users, no revenue, often not even a clear company yet.

So the question I’m increasingly focused on isn’t “Is this person brilliant?”

It’s: what kind of system will this person build once other people, capital, and power are involved?

That’s the throughline of what I care about heading into 2026.

‍

The bridge from research to building isn’t execution. It’s “ownership.”

A common mistake is assuming that the gap between research and building is about learning to execute or move faster. More often, it’s about whether someone naturally takes ownership before they’re forced to.

You can see early versions of this even in research settings:

• building tools, datasets, or collaborations without permission
• making directional bets before consensus forms
• pursuing problems that feel important even if they’re unfashionable
• narrowing optionality as conviction grows instead of hoarding it

These aren’t productivity signals. They’re incentive signals, indicators of how someone will behave once they’re responsible not just for ideas, but for other people’s time and trust.

‍

What this reveals about founder signal

I’m increasingly drawn to researcher-founders who:

• are intrinsically motivated by the problem, not just by recognition
• are comfortable acting under ambiguity without needing permission
• think early about collaborators and environments, not just outputs
• increase their exposure, reputational, financial, or career, as belief grows

These traits are hard to fake. They tend to show up long before a product exists, and they map directly to how founders later design incentives, distribute power, and ask others to commit. And in my experience, they predict whether someone will build systems that compound, or that quietly optimize for short-term validation.

I’ve seen this most clearly in a researcher who spent years working on a problem, long before starting anything new. Instead of stopping at papers or prototypes, they pushed their work into production systems with real-world consequences; systems where performance, reliability, and downstream impact actually mattered.

They took ownership of outcomes that affected other teams, accepted tradeoffs that made the work less elegant but more useful, and kept iterating even when there was no clear evaluation function or guarantee of success. Importantly, none of this was required of them and it definitely wasn’t the fastest path to recognition. It was simply the right way to make the work matter.

By the time they started something, their “founder behavior” wasn’t new. Starting a company was just the next, natural expression of it.

‍

Why this connects to incentives, intensity, and longevity

A lot of what I’m focused on in 2026 is about avoiding false positives:

• founders who sound right but design misaligned systems
• companies that move fast but burn people out
• incentives that look generous but distort behavior over time

Researcher-founders who successfully become builders tend to have an intuitive grasp of these dynamics early. They don’t just care about being right. They care about creating environments where others can move with them, stay with them, and take real ownership.

That instinct matters more than polish, speed, or even initial idea quality.

‍

A cleaner way to frame the distinction

Instead of asking whether someone is a researcher or a builder, I find it more useful to ask:

When outcomes are unclear and no one is watching, do they optimize for truth alone, or for systems that allow others to commit alongside them?

That question shows up everywhere: in how people talk about incentives, how they handle optionality, how they think about equity, and how seriously they take the long term.

Many of the most important companies of the next decade will come from deep research. But they won’t be built by research alone.

They’ll be built by people who can carry rigor into environments where incentives matter, ambiguity is constant, and ownership begins long before success is guaranteed.

That’s what I’m paying attention to as I head into 2026 and it’s shaping how I think about our major research university effort and how I evaluate founders long before there’s anything to measure.