Capacity for different types of work

You are not running out of energy — you are running out of the wrong kind

You track your deep work hours. You have a number — maybe four, maybe five on a good day — and when you hit it, you declare yourself done. Cooked. Out of capacity. You switch to low-value tasks or call it a day.

But here is something you have probably noticed without naming it: writing a strategy document and debugging a data pipeline are both deep work, and they do not drain you the same way. You can write for three hours and feel completely hollowed out, then sit down with a spreadsheet and feel your attention sharpen rather than blur. Or you can spend four hours in analytical work and feel wrung out, unable to look at another table, but perfectly capable of walking into a meeting and being socially present for the next hour.

This is not willpower fluctuation. This is not randomness. This is the signal of a fundamental fact about how your cognitive system works: you do not have one pool of capacity. You have several. They deplete at different rates. They recover at different rates. They have different daily ceilings. And if you treat them as a single pool — scheduling your day as though "work" is one homogeneous resource that gets used up uniformly — you will consistently underperform your actual capacity while feeling more exhausted than necessary.

The multi-pool model of cognitive capacity

The insight that capacity is type-specific, not monolithic, has deep roots in cognitive science. The most formal treatment comes from Christopher Wickens, whose Multiple Resource Theory, developed across several decades of research at the University of Illinois, proposes that human attention is not a single undifferentiated pool. Instead, Wickens argued that humans have multiple pools of attentional resources organized along several dimensions: the stage of processing (perception versus cognition versus response), the code of processing (verbal versus spatial), and the modality of input (visual versus auditory). Tasks that draw from the same pool compete for resources and cause interference. Tasks that draw from different pools can operate in parallel with minimal degradation.

Wickens built this theory to explain performance in complex environments — aviation, driving, industrial operations — but the implication for personal capacity planning is direct. When you write a strategy document, you are drawing heavily on verbal-cognitive resources: generating language, holding abstract concepts in working memory, making judgment calls about which ideas to include and how to sequence them. When you switch to data analysis, you shift to spatial-cognitive and procedural resources: pattern recognition in visual displays, logical sequencing, systematic comparison. The pools overlap, but they are not identical. Depleting one does not necessarily deplete the other.

For practical purposes, you can model your capacity across four broad pools that map to the major types of work most knowledge workers perform:

Creative capacity — generative, open-ended work. Writing, designing, brainstorming, strategic thinking, ideation. This draws on divergent thinking: generating multiple possibilities from a single prompt. It requires holding ambiguity and making associative leaps. It is metabolically expensive and depletes quickly. Daily ceiling: typically two to four hours.

Analytical capacity — convergent, detail-oriented work. Debugging, data analysis, proofreading, financial modeling, code review. This draws on convergent thinking: following logical steps to a correct answer. It depletes more slowly than creative capacity but is more sensitive to interruption. Daily ceiling: typically three to five hours.

Social capacity — interpersonal work. Meetings, collaboration, negotiations, difficult conversations, presentations. This draws on emotional processing, empathy, and real-time social monitoring. It is uniquely draining because it requires continuous model-building — constantly predicting what the other person thinks and adjusting accordingly. Introverts and extroverts have different ceilings, but both have ceilings. Daily ceiling: typically one to four hours.

Administrative capacity — logistical work. Email, scheduling, invoicing, filing, routine correspondence. This draws on procedural execution among low-complexity items. It depletes slowly because each task requires little cognitive investment. Daily ceiling: can often fill remaining hours without significant degradation.

The exact taxonomy matters less than the structural insight: the pools are separate. Your day is not a single eight-hour block of generic capacity. It is four parallel reservoirs, each with its own fill level, drain rate, and recovery pattern.

The science of separate pools

Wickens is not the only researcher who pointed to this structure. The evidence converges from multiple directions.

Daniel Kahneman's distinction between System 1 (fast, automatic, intuitive) and System 2 (slow, deliberate, analytical) processing, articulated most fully in Thinking, Fast and Slow (2011), captures a related dimension. System 2 processing — the effortful, conscious reasoning that underlies both creative and analytical work — draws on a limited resource that Kahneman describes as a kind of cognitive budget. But Kahneman also notes that different System 2 tasks impose different types of load. Mental arithmetic, language production, and spatial reasoning all require System 2 engagement, but they do not all deplete the same substrate in the same way. A person who is mentally fatigued from sustained verbal reasoning may still perform well on a spatial task because the specific resources demanded are partially distinct.

Stephen Monsell's research on task switching, particularly his 2003 review in Trends in Cognitive Sciences, adds another layer. Monsell demonstrated that switching costs are not uniform — they depend on the similarity between tasks. Switching between two verbal tasks (writing an email, then writing a report) incurs a higher cost than switching between a verbal task and a spatial task (writing an email, then reviewing a floor plan). The explanation is resource overlap: tasks that draw from the same pool compete during the switch, requiring more time and producing more errors. Tasks that draw from different pools can be switched between more fluidly because the reconfiguration is less extensive.

Jim Loehr and Tony Schwartz, in The Power of Full Engagement (2003), proposed a model of human energy with four separate dimensions: physical, emotional, mental, and spiritual. Built from decades of performance coaching with athletes and executives, their key observation was that performers who managed energy across all four dimensions outperformed those who focused on time management alone. A person who was physically recovered but emotionally depleted could not perform. Each dimension had its own depletion pattern, recovery requirements, and capacity ceiling.

What happens when you treat capacity as monolithic

The practical cost of ignoring multi-pool capacity is a pattern you probably recognize in your own behavior.

You schedule a full morning of creative work — writing, strategizing, ideating. By 11 AM, creative output is degrading. Sentences are flatter. Ideas are derivative. You are recycling rather than generating. But you keep pushing, because you have "deep work time" blocked until noon and the schedule says you should still be producing. You grind out another ninety minutes of mediocre creative work. Then, in the afternoon, you face an analytical task — reviewing a budget, debugging a process, analyzing data — and you cannot engage with it. Not because analytical capacity was depleted (you never used it), but because you are globally fatigued from pushing past the creative ceiling. The overshoot in one pool bled into the others.

This is the overcommitment cost from The cost of overcommitment, expressed at the pool level. Exceeding the capacity of a single pool does not just reduce the quality of that pool's output. It generates systemic fatigue — frustration, resistance, cortisol — that degrades capacity across all pools. You do not merely run out of creative juice. You poison the analytical well.

The alternative is what the product manager in the example above did instinctively: sequence across pools. When creative capacity hit its ceiling, switch to analytical. When analytical capacity hit its ceiling, switch to social. When social capacity hit its ceiling, switch to administrative. Each switch allows the previously depleted pool to begin recovering while a fresh pool takes the load. The total hours of high-quality work per day increase not because you have more capacity but because you stop wasting capacity by stacking same-type work until the pool breaks.

Mapping your personal capacity profile

Your capacity profile is unique. The four-pool model is a framework, not a prescription. Some people have six hours of creative capacity and ninety minutes of social capacity. Others have five hours of social capacity and two hours of creative capacity. Some people recover creative capacity during a lunch walk. Others need a full night of sleep before the creative pool refills. The framework tells you the pools exist. Your job is to discover your specific ceilings and depletion curves.

The mapping process requires one week of honest tracking. For each work block of thirty minutes or more, classify the type and note when quality began to degrade. Quality degradation has specific observable markers: for creative work, you start repeating yourself, ideas feel forced, and you default to the obvious rather than the novel. For analytical work, you start missing errors, losing your place in sequences, and rereading the same data without absorbing it. For social work, you disengage from conversations, respond on autopilot, and feel irritation rising at normal interpersonal friction. For administrative work, you start deferring simple tasks and making careless mistakes on routine items.

The ceiling is not the point at which you stop completely. It is the point at which quality-per-hour drops below a threshold that makes continuing worthwhile. A writer who produces 800 words of publishable prose in hour one and 200 words of deletable prose in hour four has a creative ceiling of roughly two to three hours, not four.

Once you have your ceilings, you can construct a daily schedule that respects them. The general pattern: creative work first (it requires the freshest cognitive state), analytical work second (it tolerates moderate fatigue), social work in the middle or late afternoon, and administrative work in the transitions and at the end of the day. But this pattern is a starting hypothesis. Some people are more creative in the evening. Some find social interaction energizing and use meetings to recharge between solo blocks. The only way to know your pattern is to test it, measure it, and adjust.

Pool-switching as recovery

There is a subtlety here that elevates multi-pool awareness from a scheduling tactic to a genuine capacity multiplier. When you switch from a depleted pool to a fresh one, the depleted pool does not merely stop draining — it begins actively recovering. This is not metaphor. It is consistent with the neuroscience of cognitive fatigue.

Stephen and Rachel Kaplan's Attention Restoration Theory suggests that cognitive fatigue is not a depletion of a biochemical substrate (like glycogen in muscles) but a depletion of inhibitory capacity — the ability to suppress irrelevant stimuli and maintain focus on a chosen target. Different types of work tax different inhibitory circuits. When you switch to a type of work that engages different circuits, the previously taxed circuits get a form of rest even though you are still actively working.

This means that the sequence creative-then-analytical is not just additive — it can be synergistic. The analytical work provides partial recovery for creative circuits, so that the next morning's creative capacity is higher than it would have been if you had pushed creative work for six continuous hours the day before. You are not just avoiding the overshoot penalty. You are actively accelerating recovery by using different pools in sequence.

The practical implication: do not rest when a pool runs dry. Switch. Save true rest — walking, napping, disconnecting — for when all pools are simultaneously depleted, which, if you sequence well, will happen later in the day than you expect.

Common misconfigurations

Three scheduling patterns reliably destroy multi-pool capacity:

Same-type stacking. Blocking an entire day for one type of work: "Monday is my writing day," "Wednesday is meeting day." This exhausts one pool hours before the day ends while leaving other pools completely untouched. The depletion is deep enough to impair recovery, and the unused pools represent waste.

Reactive sequencing. Letting your schedule be determined by whatever arrives in your inbox or whoever messages you next. This produces random task-type sequences that maximize switching costs (because you have no control over task similarity) and prevent you from using any pool efficiently (because you never spend enough continuous time in one type to reach peak output before being pulled to another).

Recovery avoidance. Filling every transition between major work blocks with more work of the same type because it feels productive. Recovery windows that get filled with low-grade same-type work are not recovery windows. They are depletion extenders.

The Third Brain

Classifying your work by type is straightforward for obvious cases — writing is creative, debugging is analytical, meetings are social. But much of knowledge work is hybrid. A design review meeting is social and analytical. A brainstorming session with a team is creative and social. Strategic planning is creative and analytical. When you track by hand, you approximate. When your Third Brain — an AI system with access to your task logs and calendar — classifies for you, it can assign fractional weights to each pool based on the actual cognitive demands of each task.

More importantly, an AI system can monitor depletion across pools in real time. If you log your capacity at the end of each work block (a simple 1-to-5 rating for each pool: "creative: 2, analytical: 4, social: 5, admin: 5"), the AI can track your depletion curves across days and weeks. It can detect when your creative ceiling has shifted — dropped because of sleep debt or risen because of a productive streak. It can suggest task-type sequences that optimize for total output rather than for the type of work you feel like doing in the moment, which is often the type you should not be doing because that pool is already depleted.

The AI can also surface recovery patterns invisible to your own perception — maybe your analytical capacity recovers faster on exercise days, or your social ceiling is higher when you start with solo creative work. These patterns exist in the data. Without an external system tracking them across weeks and months, they remain invisible. With one, they become schedulable.

The bridge to building capacity

You have been planning around your existing capacity — measuring it, buffering it, protecting it from overcommitment. You now understand that "existing capacity" is not one number but several, and that the relationship between those numbers determines how much total work you can do in a day without degradation.

But these ceilings are not fixed. They can be raised. A person who currently has two hours of creative capacity per day can, through deliberate practice and progressive overload, build that ceiling to three hours, then four. A person with ninety minutes of social capacity can build tolerance and skill until the ceiling rises to three hours. The pools are trainable — but only if you train them correctly, which means gradually, consistently, and with adequate recovery between training loads.

That is the subject of Building capacity gradually: building capacity over time. The multi-pool model you learned today tells you which pools to train. The next lesson tells you how.