AI Changed How I Think About Time
Listen
For most of my career I thought productivity was about managing time. I now think that's wrong. The real constraint isn't time, it's cognitive energy, and not all cognitive energy is equal. A few months ago I started tracking my energy through the day, and the pattern was consistent: my mornings weren't slightly better than my afternoons, they were a different resource. AI made this matter more, because once execution gets cheap the question stops being what to do and becomes what deserves my best energy.
A morning hour, for me, is worth about three afternoon hours. The exact multiple doesn't matter. The principle does: hours are not interchangeable, so I stopped budgeting in hours and started budgeting in units.
Why are some hours worth more than others?
Because what I can do well changes through the day. The same sixty minutes buys a different class of work depending on when I spend it.
| Time of day | What I can actually do well | Best used for |
|---|---|---|
| Morning | Think strategically, solve ambiguous problems, make tradeoffs, see patterns | Judgment |
| Afternoon | Execute, follow process, build, attend meetings | Execution |
| Evening | Learn, explore, experiment, build skills | Capability building |
The traditional model assumes every hour is the same. My experience is the opposite. So every day starts with a fixed budget, not of hours but of units. Morning is worth 3, afternoon 1, evening 2, call it 6 units total. The first mistake most people make is treating all 6 the same. Using a morning unit to approve an invoice is like using venture capital to buy coffee. The resource is too valuable for the task.
What new problem does AI create?
It removes the old excuse. Historically a lot of tasks demanded your best energy because there was no alternative. Now AI can build the report, write the code, summarize the document, stand up the dashboard, and run the workflow. The question is no longer "what should I do?" It's "what deserves my highest-quality cognitive units?" That's a harder question, and most people never ask it.
What are the three types of work?
I sort work into three categories, and each one belongs to a different part of the day.
| Type | Creates value through | Examples | When |
|---|---|---|---|
| Judgment | Thinking | Pricing, strategy, hiring, prioritization, resource allocation | Morning |
| Execution | Throughput | Building, operating, documentation, analysis, shipping | Afternoon |
| Capability building | Future capacity | Communication training, learning AI, systems design, reading, experimentation | Evening |
Judgment is where human leverage is highest. A dashboard can't make a pricing decision and an AI workflow can't set strategy, so the decisions go in the morning. Execution is implementation after the decision is made, so it goes in the afternoon. Capability building doesn't help today but raises tomorrow's ceiling, so it goes in the evening when I can't make my best calls anyway.
Why do builders get trapped?
Because AI introduces a new kind of procrastination: building leverage. A dashboard feels productive. An automation feels productive. A workflow feels productive. Sometimes they are. But it's easy to confuse building leverage with creating value, and a system is only valuable if it supports a future decision or action.
I've fallen into this many times. I sit down to solve a pricing problem. Three hours later I've improved the dashboard. The dashboard got better. The pricing decision didn't. That's a morning unit spent on something an afternoon unit could have done, while the thing only the morning unit could do went unanswered.
What is the real job, then?
Allocating cognitive capital, not building the most systems. The highest-leverage people aren't the ones who ship the most automations. They're the ones who consistently point their best units at the highest-return work. Every morning is an investment committee meeting. Multiple projects compete for scarce capital, and each one has to justify the unit, not because it's interesting or comfortable, but because it has the highest expected return.
I weight each project on four questions:
- Impact. If it works, how much value does it create?
- Repeatability. How often will the benefit recur?
- Persistence. Will it still matter in six months?
- Relative priority. What else could I do with this unit? This is the one that matters most. A project doesn't get worse because it changed. It gets worse because a better opportunity showed up next to it.
What's the new bottleneck?
Deciding where your best energy goes. AI is making execution cheaper every month, so the ability to build is becoming abundant and the ability to allocate attention is becoming scarce. Not everyone will have the best tools, prompts, or automations, but the people who consistently spend their best cognitive units on the highest-return work will outperform the ones who simply stay busy. The bottleneck is no longer execution. It's allocation.
Frequently asked questions
How do you know which hours are your best? Track your energy for a couple of weeks and watch what kind of work feels easy versus forced. Most people find one block where strategic thinking is effortless and another where only execution is realistic. Protect the first block and stop scheduling decisions in the second.
Isn't this just "do hard work in the morning"? It's more specific. The point isn't difficulty, it's type. Reserve your peak units for judgment, the work only you can do, and push execution and learning to the hours where AI or lower-energy effort is good enough.
Why is building systems risky if AI makes it easy? Because building feels productive whether or not it serves a real decision. The danger is spending peak energy improving a tool while the decision the tool was meant to support stays unresolved. Build only when a future action depends on it.