Being able to see someone else's good idea immediately seems like an unquestionable advance.
No need to go and ask. No need to wait for a meeting. The shared screen holds the best current proposal. When in doubt, return to it.
But is everyone having the same place to return to the same as a group being able to search farther afield?
Three people, twenty-five cities
In a study published in 2018 by Ethan Bernstein, Jesse Shore, and David Lazer, anonymous groups of three searched for short routes through 25 cities. One condition let participants see the others' solutions every round, another every third round, and a third not at all. The researchers varied not just whether people were connected, but the interval between connections.[1]
Without a feature that restored a participant's best past solution at a click, the share of trials finding an optimal solution was 33.3% with constant sharing, 48.3% with intermittent sharing, and 44.1% with no sharing. The difference between intermittent and constant sharing was significant after adjustment; the difference between intermittent and no sharing was not. For average solution quality, however, no significant difference was found between intermittent and constant sharing, while intermittent sharing performed better than no sharing.
This is not a story in which thinking alone wins. Nor is it one in which thinking together wins. The interesting condition alternated between receiving good solutions and searching for different ones.
There are limits. This was a short route-search task, not free discussion or collaborative editing. Participants viewed other people's solutions. The study did not show that workplace teams, or people working with AI, would perform well at the same intervals.
Being helped, and becoming alike
The result raises a question less about how often collaboration happens than about what happens after it.
Seeing someone else's proposal gives us a foothold. We no longer have to create a starting point from nothing. The energy saved may carry us farther. But a foothold can also become a meeting place. If everyone gathers there and begins making small adjustments, other places can quietly disappear from consideration.
Not because the first idea was bad. Because it was good enough that leaving it no longer seemed necessary.
Improving the quality of proposals will not necessarily solve this difficulty. A rough idea makes it easy to suggest a different direction. A polished idea invites an equally polished critique. Its surroundings become increasingly well explored. Beyond them, no one has thought to ask.
This next point is an editorial reflection beyond the study: the question seems worth placing in a setting where AI supplies a polished proposal from the outset. The more useful the first answer is, the more we might ask whether making revisions to it has become the whole of thinking.
That does not mean we should decide to withhold answers. Thinking alone can leave us stuck in the same place. Encountering someone else's idea may be precisely what reveals another direction. Taking some distance is not the same as becoming isolated.
Time without meetings has a role too
“Let's think separately and bring our ideas together” can be treated as ordinary meeting preparation. From another angle, it preserves a period when people have not yet seen the same proposal. It is working time that allows them to bring back different things.
Of course, if everyone clings to their own idea and none can be adopted, that is a problem too. If independence deserves a budget, so does coming together. Making separation a virtue is as reductive as making interaction a virtue.
One interesting feature of the study is its separation of a group's average performance from the best solution the group found. Helping everyone get unstuck and enabling someone to discover a new exit are both desirable in an organization. They may not be maximized by the same mechanism.
In daily operations, reducing variation can be valuable. When we are still searching for what to make, allowing variation for a while may matter too. Before diagnosing “not enough collaboration,” we might ask whether what is missing is sharing, or the differences that exist before sharing.
The study also found that adding the ability to restore a previous best solution weakened the benefit of intermittent sharing. That does not make saving or record-keeping itself a villain. Instead, it leaves another question: when there is somewhere to return to, how do we support continuing to search rather than returning?
We need not delete the history. We might write an idea that has not yet appeared before opening it. Such small practices are not prescriptions tested by this experiment, but they can become hypotheses to try in our own settings. Rather than borrowing a number of minutes apart from a paper, it may be more useful to look at what has been added when people return.
Sometimes nothing emerges while people are apart. Demand a new idea from every absence, though, and independent time becomes another reporting meeting. What we need is not faith that separation always makes us creative, but an understanding that showing one another everything, all the time, is not the only design option.
Trusting someone can also take the form of not immediately looking at their work in progress. Something not yet put into words, something that would demand explanation if shown too soon, may need a little time to grow. That waiting time does not always produce results. But the speed of making something shareable is not the only speed of exploration.
When we come together again, we may not need to report every intermediate thought. A changed assumption, a new question, a lingering snag: the smallness of what we bring back and its power to change the next discussion are separate matters.
If we call only connected time collaboration, we push the time that enriches connection outside collaboration altogether.
When we next gather, will we bring continuations of the same answer? Or paths that no one else has yet seen?