Jay's desk was a cockpit. Three monitors arranged in a gentle arc, each one assigned a permanent role that never changed. The left monitor showed dashboards—agent activity, scenario completion rates, satisfaction metrics, token consumption. The middle monitor was for code, or more precisely, for the specs and scenarios that generated code. The right monitor was for Hacker News.
The right monitor was always on. It displayed the HN front page in a browser tab that auto-refreshed every five minutes. Sometimes Jay would glance at it mid-sentence, catch a headline, file it away, and continue talking without losing his thread. The right monitor was ambient. It was background radiation from the internet's id, a constant low-frequency signal of what the technical world was thinking about at any given moment.
Navan's desk was a monastery. One monitor. One laptop. One notebook.
The monitor displayed whatever Navan was actively working on—twin code, scenario specs, Block Kit rendering tests, swift-gopher updates. It showed exactly one thing at a time. When Navan switched tasks, the entire monitor changed. There was no ambient feed, no peripheral awareness, no background signal. There was the thing he was doing, and there was everything else, and the everything else didn't get a screen.
The notebook sat to the right of the keyboard, always open, always within reach. When Navan needed to think about something that wasn't on the screen, he picked up his pen and wrote. The notebook was his second monitor, except it had no backlight and no refresh rate and it never showed him anything he hadn't put there himself.
Jay noticed the contrast in September. He'd walked over to ask Navan about a Slack twin rendering issue and found Navan's single monitor displaying nothing but a terminal with green text on a black background. The notebook was open to a page covered in a diagram that looked like a state machine.
"How do you work with one monitor?" Jay asked. It was a genuine question, not a judgment.
"How do you work with three?" Navan replied. Also genuine.
"I need to see everything at once. The dashboards tell me what's happening. The code tells me why. HN tells me what's next."
"I need to see one thing at a time. Everything else is noise until I need it."
"Doesn't that slow you down?"
"Doesn't that distract you?"
They looked at each other and both realized the answer to both questions was no. Two people, two cognitive architectures, two workspace configurations that were perfectly optimized for the mind that used them. Jay's three monitors were a distributed system. Navan's one monitor and notebook were a single-threaded process with a write-ahead log.
Neither configuration was superior. They were homologous—structurally different, functionally equivalent. Both produced the same output: good specs, good scenarios, good thinking about complex systems. The input medium didn't determine the output quality. The mind did.
Jay went back to his three screens. Navan went back to his one. The agents, indifferent to either arrangement, continued their work.
Navan's notebook as "a second monitor that never showed him anything he hadn't put there himself." That's the most precise description of why analog tools work that I've ever read. No notifications. No refresh. Just your own thoughts.