There used to be a natural pace to work that came from the friction embedded in doing it. Drafting something took time, so a request for a draft naturally sat in a queue for a while before it could be addressed. Research took time, so a question requiring research had a built-in delay before an answer could arrive. This friction wasn't designed as a prioritization system, but it functioned as one anyway: things that required more effort naturally got triaged and sequenced, because they couldn't all happen immediately. AI has removed a significant amount of that friction, and in doing so, it's removed the informal prioritization system that used to come with it. Nearly everything can now be actioned immediately. And immediate actionability is quietly getting mistaken for immediate necessity, in a pattern that's driving a specific and underexamined form of overwhelm. ------------- Context ------------- Before AI, the time required to complete a task functioned as a natural filter on what could realistically happen right now versus what had to wait. A request that would take three hours to fulfill couldn't be actioned in the next ten minutes, regardless of how urgently it was framed, simply because the work took time. This created an implicit form of triage: things got sequenced by a combination of actual priority and practical feasibility, and the feasibility constraint did a lot of quiet work in keeping the pace of a day manageable. AI has collapsed the feasibility constraint for a huge range of tasks. A request that used to require hours can now be actioned in minutes. This is a genuine advantage in many cases. But it also means that the natural pacing mechanism that used to exist alongside the feasibility constraint is gone, and nothing has automatically replaced it. Everything that arrives now carries an implicit invitation to be handled immediately, because immediate handling is now technically possible in a way it never used to be. The psychological effect of this shift is significant and underappreciated. When something is technically actionable right now, there's a pull toward treating it as though it should be actioned right now, even when the actual priority of the task hasn't changed at all. Feasibility and urgency are different things, but in a world where almost everything has become instantly feasible, the distinction is easy to lose.

AI has compressed the time required for most work that's documented and explainable: work where the process, the standards, and the reasoning can be captured and communicated clearly. What AI hasn't touched, and can't help with, is work that depends entirely on knowledge that exists only in someone's head and has never been written down anywhere. This creates an increasingly stark and underexamined divide inside most businesses. The documented, explainable work is getting dramatically faster. The undocumented, tacit knowledge is becoming, by comparison, a disproportionate bottleneck and a genuine point of fragility, because it's the one category of work that AI adoption does nothing to address until someone takes the separate step of actually capturing it. ------------- Context ------------- Every business accumulates tacit knowledge over time: the specific reasons a particular client relationship requires careful handling, the informal workaround for a recurring operational problem, the judgment calls a founder makes intuitively that have never been articulated as an explicit process, the history behind why something is done a certain way. This knowledge was always somewhat risky to keep undocumented, but for a long time, the risk was manageable because most work moved at a pace where the person holding the knowledge was usually available when it was needed. AI adoption changes the risk calculation significantly, for two connected reasons. First, as documented work gets dramatically faster, the undocumented work becomes a proportionally larger share of total bottleneck time, simply because everything around it has sped up while it hasn't moved at all. Second, and more subtly, businesses that are scaling their output using AI are often taking on more volume, more clients, more complexity, faster than before, which increases the number of situations where tacit knowledge would be needed and decreases the amount of time available to informally transfer it the way it might have been transferred in a slower-moving business.