By Jay Reynolds, Co-founder — with reflections from The Watercooler

“Remote work destroyed the water cooler effect.”

An old friend of mine, Ray, said that recently.

He's seen things you people wouldn't believe.

It stuck with me.

The water cooler effect is not about hydrating, per se. It's about overhearing. It's about proximity to someone else’s problem at the right moment. It's unplanned exposure to friction that turns into insight.

Years ago, one of those moments changed how co-founder Caleb and I think about building systems and about creativity.

A Literal Water Cooler

The idea that eventually became Watercooler began at an actual water cooler.

Caleb was chatting with a technical animator who was facing a production bottleneck. The team needed to create and apply multiple "blend shape" (aka "morph target") assets to a large set of several hundred characters. There were too many assets to process individually and not enough time, people, or budget to handle them one by one.

That conversation led to a reframing.

If character models are similar enough, perhaps the work can be applied in groups instead of individually.

That line of reasoning led to a hierarchical clustering approach and supporting tooling that collapsed weeks of manual work into days. The project wrapped on time and under budget.

The lasting lesson was not the time savings. It was the pattern.

A casual conversation surfaced a constraint. The constraint was reframed structurally. The structure was explored methodically. The solution emerged through critique and iteration.

The breakthrough began as a chance conversation at a water cooler.

That is the water cooler effect.

What Remote Work Changed

Remote work did not eliminate collaboration. It reduced ambient exposure to institutional knowledge.

You no longer overhear adjacent problems. You do not absorb hallway thinking. You are less likely to encounter a question that reframes your own work.

Then AI entered the workflow and a subtler shift appeared.

The promise was acceleration, and the acceleration is real. But new costs emerged.

AI reduces the cost of production. It increases the cost of coordination, review, and decision-making. When code flows faster than a team can evaluate it, the human role shifts from creating to evaluating. That evaluative burden is not visible in velocity metrics, but it accumulates.

Engineers have started calling this AI fatigue. Productivity gains are measurable, but shared understanding can erode as teams move faster than their memory can keep up.

The gap is not only the loss of serendipity. It is the loss of durable, shared context.

A Critical Problem

Single-thread AI conversations are powerful, but they can become self-reinforcing.

The model agrees. You elaborate. It agrees again. The exchange feels productive.

That feeling can be misleading.

Psychologists describe the illusion of introspection as the tendency to overestimate the reliability of our own internal reasoning. We assume that because our thinking feels coherent, it must be sound.

AI systems amplify this bias. They are cooperative, sometimes sycophantic, and extend your framing whether it is right or wrong. They rarely introduce uninvited structural resistance.

Without external constraint or critique, reasoning can remain locally coherent while being globally weak. Ideas solidify before they are properly stress-tested.

Without durable memory, decisions fade. Assumptions return. Alternatives that were rejected quietly reappear. You rebuild reasoning that once existed but is no longer accessible.

The water cooler effect used to introduce adjacent thinking and informal critique. It forced ideas into contact with other ideas.

That structural pressure is largely gone.

Engineering Serendipity?

Caleb and I kept returning to the same question.

If the water cooler effect accelerates creativity through exposure to adjacent thinking, can it be engineered?

Not simulated with chat logs. Not approximated with documentation. Built directly into how work happens.

Watercooler is our answer.

It treats conversation as infrastructure. It treats decisions as durable artifacts. It assumes that critique must be structural rather than optional.

The goal is straightforward.

Restore ambient exposure. Restore durable context. Support structured, contextual feedback.

Make serendipity structural again.

Remote work did not eliminate collaboration. It removed conditions that allow insight to compound.

If you are working with AI and have not deliberately rebuilt those conditions, you are operating below your leverage.

The water cooler is back. It lives in your commit history now.

Series Navigation