Chapter 3

The Attention Recession

Reclaiming Sustained Focus

“Concentration is the secret of strength.”

— Ralph Waldo Emerson

Gloria Mark has spent twenty years doing something almost nobody else does: following knowledge workers around with stopwatches, logging exactly where their attention goes. Her findings trace a collapse. In 2004, the average worker held attention on one screen for about two and a half minutes before switching. By the time of her recent studies, the average was forty-seven seconds. And her research with colleagues at Humboldt University found that once genuinely interrupted, workers took on the order of twenty-three minutes to fully return to the task — compensating by working faster, at a measurable cost in stress.

Those numbers describe the world before AI assistants. What AI adds is stranger and harder to see, because it does not interrupt you at all.

The Eroder That Never Interrupts

Social media steals attention loudly. AI erodes it politely — by assisting. Every time a tool absorbs the demanding middle of a cognitive task, the requirement to sustain focus drops. The work still gets done; what disappears is the practice. And sustained attention is a capacity, not a preference: it strengthens with use and atrophies without it, exactly like the hippocampi of the drivers who stopped navigating. A professional who spends two years letting tools carry every hard stretch of thinking does not merely get out of the habit of deep focus. They lose the ability, and discover the loss the first time a problem refuses to be delegated.

Attention deserves this early place in the book for one reason: it is the medium every other form of depth is built in. You cannot develop genuine expertise in a domain you cannot attend to for sustained periods. You cannot build real relationships with people you cannot be fully present with. Lose the medium and you lose everything downstream of it.

Deep attention is not a productivity technique. It is the medium through which all genuine expertise is built.

The Morning the Ideas Came Back

A product design team I learned of ran a simple protocol: two hours every morning with no AI tools, no messages, no meetings — just the hardest design problems and their own minds. The first two weeks were uncomfortable in the way an unused muscle is uncomfortable. By week six, the quality of ideas from the morning sessions had visibly improved — and, more telling, the designers reported holding more complexity in mind during the afternoon sessions that did use AI. The capacity built in the constrained hours was transferring to the augmented ones.

That transfer is the whole argument of this chapter. Protecting attention is not a retreat from the tools. It is what keeps the mind directing the tools strong enough to be worth augmenting.

The Attention Inventory

For one week, track two numbers honestly: hours spent in sustained, uninterrupted focus on one demanding task, and the number of times you reached for an AI tool before genuinely engaging with a problem. Most professionals who run this audit are startled by both. The second number is your personal attention recession index — and unlike Mark’s statistics, it is entirely within your control.

The Multitasker’s Paradox

It gets worse before it gets better, and the research on heavy multitasking explains why willpower alone won’t fix a fragmented day. In a widely cited 2009 Stanford study, Eyal Ophir, Clifford Nass, and Anthony Wagner compared people who habitually juggle many streams of media against those who don’t — expecting the multitaskers to be better at filtering distraction, since they had so much practice. They found the opposite: heavy multitaskers performed worse at ignoring irrelevant information, worse at task-switching, and worse at holding information in working memory. Practice at fragmentation did not build a skill. It built a deficit, and a deeper one than occasional multitasking produces.

The uncomfortable implication is that attention doesn’t partially recover between tasks the way rest recovers a muscle. It compounds in the wrong direction, the same way every capacity in this book does when starved of practice. Which means the daily deep-work block below is not a nice-to-have alongside a fragmented schedule. For anyone who already multitasks heavily, it is closer to rehabilitation — and the Stanford data suggests it is worth starting before the fragmentation becomes the only mode the mind knows.

Depth in Practice: Winning the Morning in an Interrupt-Driven Office

The ninety-minute block fails for most people not from lack of will but from lack of protocol — so here is one that survives contact with a real office. The night before, choose the one problem the block belongs to and write it on a card; deciding in the morning burns the freshest attention on triage. The block itself goes on the shared calendar with a name — not ‘busy,’ which invites negotiation, but ‘deep work: pricing model,’ which announces a commitment. Status message on, phone in a drawer in another room — the last chapter’s research on mere presence applies to your own devices most of all.

Then the piece almost everyone skips: the emergency contract. Tell the three people most likely to interrupt you exactly how to reach you if something genuinely cannot wait — a phone call, twice in a row. In six weeks of running this protocol, most professionals discover the contract is invoked roughly never, which quietly reveals how few of the day’s interruptions were emergencies wearing urgent clothing.

Expect the first two weeks to be uncomfortable — the design team from earlier in this chapter reported exactly that — and expect the reflex to check something to fire every few minutes. Note it on the card each time it fires, and keep working. The tally will fall week by week, and that falling number is the attention recession, in your own handwriting, reversing.

The Practice

1. Establish a daily block of at least ninety minutes for deep work: no AI tools, no notifications, no parallel tasks. Start at forty-five if ninety feels impossible, and extend as the capacity rebuilds. Defend it on your calendar like a client meeting.

2. Identify the three problems in your current work that most deserve your deepest attention, and give them the protected block — not the leftover fragments of the day.

3. Run the one-week inventory above. Whatever the reach-for-the-tool number is, cut it in half the following week by imposing a twenty-minute engagement rule before any prompt.