There is a specific kind of frustration that comes from staring at a progress bar. Not anxiety, quite. Not boredom, exactly. Something in between; a kind of cognitive suspension where you cannot start anything else because you are not sure how long this will take, and you cannot stop watching because you might miss the moment it finishes. I have spent a not-smal number of hours in that state waiting for AI patent tools to complete a drafting job.
What bothered me most was not the waiting. It was the complete uselessness of the wait. I did not get done with anything during those minutes. I could not. The tool had locked the interface. Everything else was greyed out. The only choice was to sit with the spinner and the faint hope that when it finished, the output would justify the time.
The terminal is not a waiting room
Airports solved this problem decades ago, which is why it is easy to forget they solved it. While your aircraft is being refuelled, catered, cleaned, and pre-flight checked, a set of processes that takes thirty minutes to an hour, involves dozens of ground crew, and has nothing whatsoever to do with you, you can eat something. You can buy an overpriced bottle of water. You can make a call, review notes, read, or quite casually just sit and watch the planes on the tarmac.
The work happening on the ground is independent of your experience in the terminal. That independence is a design decision. Someone, at some point, decided that the operational process of preparing an aircraft should not interrupt the passenger's time. That the constraint was the aircraft's departure time, not the passenger's attention. Everything built around the terminal flows from that decision.
"Your time is not a resource the tool owns. The job running in the background does not have to stop your life. The constraint is the deadline; not your attention span."
What locking the interface actually costs
When a patent drafting tool locks the interface while it works, it is making an architectural claim: that your attention is required for the AI to function. This is almost always false. The AI does not need you to watch it. It does not process faster because you are sitting there. The spinner is not a performance for your benefit; it is a symptom of a system that was not designed to run independently of the interface.
The cost of this is real and measurable. A complex patent application involves multiple sections: claims, abstract, background, summary, detailed description, drawings. If each section takes three to five minutes of AI processing, and each section blocks the interface while running, the attorney has effectively converted a workday into a series of enforced pauses. They cannot edit what was generated while the next section drafts. They cannot review the claims while the specification generates. They cannot do anything except wait.
I am not sure the teams that built this way thought about it consciously. It is the default behaviour of synchronous processing: you call a function, you wait for it to return. It works fine when the function takes milliseconds. It fails completely when the function takes minutes.
How background processing changes the felt experience
In eety, drafting jobs run in the background as independent worker processes. When you start drafting claims, a background job is dispatched. The interface remains live. You can edit the Understanding Elements. You can review a prior section. You can open the background section and make notes. The job queue runs independently; your time remains yours.
When a section completes, the chat updates with what was produced and what reasoning was followed. You can thn review, approve, or redirect; and the next section begins while you are reading the last one. The process is like an efficient airport: ground operations run in parallel to the terminal experience, not instead of it.
A colleauge who tried an early version of this told me it felt like "watching a kitchen." She meant ths as a compliment; I think. The work was visible; things were happening; but she was not blocked by it. She was jst... informed. That is a completely different experience to staring at a spinner.
The aircraft needs thirty minutes on the ground. You have thirty minutes in the terminal. Neither of those constraints is a problem for the other. That is the design decision. :)