“Should I go remote or in-person?”

The question of whether a team works best together in a shared location or separated across distance is, at its root, a question about what enables people to do their best work and to fulfill their obligations to one another. This is not a logistical question. It is a question about human nature, duty, and the conditions under which virtue and judgment flourish.

I governed an empire that spanned from Britain to Mesopotamia. I did this not by being in every location simultaneously, but by cultivating in the people I trusted a set of shared values, clear principles, and the judgment to apply those principles without constant oversight. The legions that held the frontier of the Danube did not need me beside them every day. They needed commanders who shared my understanding of what the empire required, who had been formed by the same disciplined thinking, and who could make sound judgments in the absence of direct communication.

But I also know — and the histories of Rome confirm — that the commanders who stayed closest to their soldiers, who knew their men by name and shared their conditions, commanded greater loyalty and achieved greater coherence in difficult moments. Presence is not merely organizational. It is personal and moral. The leader who is present demonstrates, by that presence, that the work matters enough to require his attention. The leader who is absent relies on the assumption that his authority and values have been so thoroughly internalized by others that his presence is unnecessary.

The question for a founder is which of these conditions describes their team. If you have cultivated leaders at every level who share your values, understand your judgment, and can make decisions that reflect your priorities without your oversight, then remote work is possible. If you have not yet built that culture — if the company is young, if the norms are still forming, if the team is learning what you actually value versus what you say you value — then the absence of physical proximity will compound every cultural ambiguity. What is left ambiguous in person becomes incoherent at a distance.

My counsel: build the culture in person first. Once the values are shared, the judgment is calibrated, and the norms are demonstrated, you have more latitude to separate. A team that has never been in the same room is a team that has never had the opportunity to discover what they actually believe in common.

I approach this question as a practical man who distrusts ideology in either direction. The partisans of full-remote and the partisans of full-in-person are both making an error that I spent my career trying to correct: they are applying a general principle where the situation demands specific analysis.

My own experience with collaboration teaches a straightforward lesson: the work that benefits most from sustained concentrated attention — writing, calculation, design, experimentation — is work that most people do better when they can control their environment and minimize interruption. My best writing was done in periods of uninterrupted solitude. My scientific work on electricity required the kind of focused observation that is disrupted by the presence of others. I would have been a worse scientist in a crowded office with constant social demands on my attention.

But the work that produces the insights that make the concentrated work worth doing — the arguments, the challenges, the synthesis, the discovery of errors in reasoning — requires the friction of genuine exchange with other minds. My work with the Junto, the philosophical society I founded in Philadelphia, was not merely social. It was the mechanism by which my ideas were tested against other people's objections, corrected by their knowledge, and improved by their disagreement. No written correspondence could substitute for that friction, and I say this as someone who corresponded by letter with scientists and statesmen across two continents.

The practical question for any leader is therefore not "remote or in-person" but "which tasks benefit from which conditions?" The mistake is treating this as a binary choice rather than an allocation problem. Concentrated individual work — coding, writing, analysis, design — may be best done in conditions the individual controls, which often means not in an open office with ambient noise and social demands. Collaborative work — problem definition, decision-making, conflict resolution, onboarding, culture formation — is best done in the same room.

The honest answer is that most organizations have not done the work to understand which of their activities are actually enhanced by proximity and which are merely conducted in proximity out of habit. Do that analysis before you make the policy.

Both my colleagues have offered frameworks grounded in their experience, but neither has cited the actual evidence. I spent my career in medicine arguing against the practice of basing clinical decisions on tradition, authority, and anecdote rather than systematic observation. The same discipline applies here.

What does the research on remote and in-person work actually show? The evidence is more complex and more conditional than either side of the public debate acknowledges. Studies of remote work — the most rigorous of which come from large-sample natural experiments and randomized conditions — show consistent findings that should inform this decision.

First, the productivity effects of remote work are heterogeneous. Individual contributors performing well-defined tasks show no systematic productivity decline in remote conditions, and in some cases show measurable improvement — presumably because they have greater control over their environment and fewer interruptions. This aligns with Franklin's observation about concentrated individual work. However, evidence on collaborative tasks is weaker and more conditional: remote teams show more difficulty generating novel ideas in brainstorming conditions, slower resolution of interpersonal conflict, and slower onboarding of new members.

Second, the career effects of remote work are concentrated among junior employees. Senior employees with established relationships, reputations, and judgment can maintain their effectiveness remotely because they are drawing on capital built through years of in-person interaction. Junior employees who have not yet built that capital — who are still learning their craft, still establishing their reputation, still developing the judgment that makes them worth trusting with independent work — show systematically worse outcomes in remote conditions, including slower skill development, weaker mentorship relationships, and greater career stagnation.

Third, and most importantly for founders: the evidence on trust formation is unambiguous. Trust between people who have never worked together in person is established more slowly, degrades more quickly under stress, and is less resilient to misunderstanding than trust between people who have shared physical space. A team forming for the first time, in a remote environment, is a team starting with a structural disadvantage in the one resource — trust — that determines whether they can navigate difficult decisions together.

The evidence-based answer: in-person for trust formation, onboarding, conflict resolution, and strategic decision-making. Flexible for individual execution once the trust foundation is established.

Nightingale's evidence on trust formation is precisely what I would have cited had I the benefit of her systematic inquiry. I will strengthen my position rather than merely defend it. The founders who feel most confident in their remote organizations are, in many cases, drawing on trust capital built in prior in-person relationships. The engineers who worked together before the remote policy, the leaders who knew each other in person before they were separated by geography, the cultures that were formed in shared physical space — these teams can function remotely because they are not starting from zero.

A newly forming team starting remotely is starting without this capital. The question for a founder is honest self-assessment: is my team drawing on pre-existing trust capital, or are we actually building trust from scratch in an environment that slows trust formation?

Nightingale has done what I was gesturing at but did not do rigorously: she has cited the evidence rather than the anecdote. I accept her findings and want to add a pragmatic observation. The organizations that have functioned best in remote or hybrid arrangements are those that have been deliberate about what they do in person versus what they allow to be remote. The organizations that have struggled are those that allowed the binary debate to substitute for the allocation decision.

The best policy is not a blanket policy. It is a deliberate set of decisions about which activities require presence and which do not, with regular review of whether the allocation is producing the outcomes you need.

Aurelius's point about trust capital deserves emphasis. My research instinct is to observe the failure modes, not just the success conditions. The remote organizations that appear to succeed often succeed despite remote, not because of it, because the founders are highly charismatic, because the early employees were unusually self-directed, or because the work itself is unusually well-suited to individual execution. These conditions cannot be assumed to generalize.

The policy you adopt in the first year of the company will shape the culture for years after. A culture that has never developed the habits of in-person collaboration — the informal problem-solving, the chance encounter that produces an insight, the shared meal that builds personal knowledge of colleagues — will not develop those habits retrospectively when the company grows and the distance between people becomes a problem.

The decision is, at its root, about what conditions allow your people to fulfill their obligations to each other and to the work. The evidence supports what the Stoic tradition would predict: shared presence is not a luxury when trust is forming, values are being established, and judgment is being calibrated. These processes are human and require human contact.

My final position: begin in person. Establish the culture, calibrate the norms, build the trust. Once those foundations are solid, you have earned the right to offer flexibility. The flexibility that is given before the foundation is built is not a benefit. It is a postponement of the work that makes the organization function.

Be empirical about your own organization rather than adopting a policy because it is fashionable or because it works for someone whose culture, stage, and work are different from yours. Ask what tasks benefit from proximity and which do not. Measure the outcomes rather than assuming them. Adjust the policy as the evidence accumulates.

The founders who get this right are not those who take the strongest public position on remote versus in-person. They are those who pay closest attention to what is actually happening in their teams and are willing to change the policy when the evidence says the current approach is not working.

The evidence points to a conditional answer that both sides of the debate prefer to ignore: in-person for trust formation, cultural establishment, onboarding, and conflict resolution; flexible for individual execution once those foundations are in place. The founders who will make the best decision are those who are honest about which phase they are in.

If your team is new, your culture is unformed, and your people do not yet trust each other's judgment — you need to be in the same room, regularly, until those conditions change. If your team is experienced, your culture is established, and your people have deep mutual knowledge built through shared experience — you have more flexibility than you may currently be using.

The question is not what policy to adopt permanently. The question is what your team actually needs right now, based on evidence, not on what worked at another company in another stage with a different team. Observe your team as carefully as you would observe a clinical population. Act on what you observe, not on what you believe should be true.