🌧 Tier 2 — Real Fatigue

Real Fatigue Is Setting In

Q: How do I know if my fatigue is getting worse vs. getting better?

Watch for escalation signals: you stop reading code you did not write with AI, you cannot explain your own recent PRs, you avoid implementation tasks, or you feel nothing when something you built works. If several are true, your fatigue is deepening — act now.

You scored 8–11 on the AI Fatigue Quiz

This is not the yellow-light warning of Tier 1. Something has shifted. Your relationship with your own code has changed, and you know it. This page is for you.

What this actually means

You are not imagining it. Something real has changed in how you work and how you feel about your work. Tier 2 is not just awareness — it is functional change. The code you ship does not feel like yours the way it used to. Problems that used to be interesting now feel opaque. You can get things done, but the satisfaction is gone.

This is more common than anyone in the industry will admit. Among engineers who score in Tier 2, the most common themes are: a growing distance from their own code, a creeping inability to trust their own output, and a quiet fear that they are falling behind in ways that will be hard to reverse.

The good news: Tier 2 is recoverable. But not by hoping it will pass. It requires intentional practice.

What is happening: Four root causes

At Tier 2, the fatigue is not random. It follows a pattern. Four forces are compounding simultaneously:

Root Cause 1

The Explanation Gap Has Widened

At Tier 1, you noticed you were accepting AI output without fully understanding it. At Tier 2, this is now a habit. The gap between what you can explain and what you are responsible for has grown wide enough that you feel it every day. You review AI-generated code and sense something is off, but you cannot always articulate why.

Root Cause 2

Identity Erosion Has Begun

Your sense of yourself as a capable engineer — someone who can look at a problem and figure it out — has quietly weakened. You used to trust your instincts. Now you second-guess them. You used to feel ownership over your work. Now it feels like you are shepherding AI output through a process rather than building something.

Root Cause 3

The Sunday Dread Is a Pattern, Not a One-Off

The Sunday evening anxiety that started in Tier 1 has become regular. It is not just “I do not want to go back to work.” It is a specific dread about not knowing whether you can do what is being asked of you — not because you lack time, but because the work no longer feels like yours to do.

Root Cause 4

Skill Drift Has Become Measurable

You have noticed it in specific moments: problems you used to solve in minutes now take longer. Code you used to write fluently now requires more starts and stops. Debugging feels murkier. You are not losing your intelligence — you are losing the specific cognitive patterns that coding used to strengthen.

What to do right now: Four Tier 2 practices

Tier 1 practices are not enough at Tier 2. You need more. These four practices are specifically designed for engineers in your position:

The Explanation Requirement (Upgraded)

Before you merge any AI-generated code — any of it — you must be able to explain three things in plain English: why this approach was chosen, what would break if a key parameter changed, and how this fits the broader system. If you cannot explain all three, do not merge it yet. Trace through it manually. Ask AI to walk you through it, then close the tab and explain it yourself from memory.

One No-AI Morning Per Week

Pick one morning — Tuesday or Wednesday tends to work best, mid-week when the rhythm is established — where you do not use AI for any implementation task. Not for bugs, not for refactors, not for “just checking.” This is a deliberate skill-maintenance practice. Start with 90 minutes. You will be slower. That is the point. The friction is the training.

The Rebuild Test (Monthly)

Once a month, take something you built with heavy AI assistance — something you shipped in the last four to six weeks — and try to rebuild a simplified version of it from scratch. No AI. No reference code. Just you, a blank editor, and what you remember. Notice what is hard. That difficulty is a map of what you have stopped learning. It is also recoverable — but only if you look at the map.

The Sunday Night Check-In

Sunday evenings, before the anxiety kicks in, write three answers in a journal or notes app: What did I work on this week that I could explain from scratch? What am I uncertain about? What do I want to understand better by next Sunday? This creates a record of your relationship with your own work that you can look back on and learn from.

What not to do at Tier 2

✗

Take a vacation and expect it to fix things. Rest helps, but the fatigue is not from being overworked — it is from a specific pattern of tool use. Without practice changes, you will return from vacation and feel the same within two weeks.

✗

Go cold turkey and ban all AI from your workflow. This is as unhelpful as unlimited use. The issue is not AI — it is the relationship you have with AI. A sudden ban without a transition plan creates anxiety and does not address the underlying habits.

✗

Tell yourself it is imposter syndrome and wait for it to pass. Tier 2 fatigue has a real, functional cause. It is not a cognitive distortion — it is a documented pattern of cognitive offloading and skill drift. Calling it imposter syndrome delays the real response, which is practice change.

✗

Redouble your AI usage to catch up. When fatigue deepens, the instinct is to work harder at the same patterns. Using more AI to solve the problem caused by AI is the definition of the trap you are already in.

What a Tier 2 recovery week looks like

Recovery at Tier 2 is not a dramatic life overhaul. It is small, consistent practices that gradually shift your relationship with AI tools and with your own craft:

Monday

Morning: look at the code you wrote last week. Can you explain three key decisions in it without looking at the AI's output? Note what you can and cannot explain.

Tuesday

Morning — 90-minute no-AI block: work on implementation without AI. Start with something small — a data structure, a utility function, a bug you have been putting off.

Wednesday

Before any AI use, apply the Explanation Requirement to your first task. Write the three answers down before you accept any AI output.

Thursday

Review your week’s AI usage. How much did you use? How much did you understand? Where was the gap the largest?

Friday

Ship something you understand completely. Not the most important feature. Not the hardest problem. Something small where you can feel what it is like to be the author, not the approver.

Sunday

Write the three answers: What could I explain this week? What am I uncertain about? What do I want to understand better?

Frequently asked questions

Is Tier 2 AI fatigue the same as burnout?

No — burnout is occupational exhaustion that accumulates over months or years. AI fatigue at Tier 2 is a more acute functional change driven by tool overuse, skill drift, and identity erosion. They can compound each other, but the recovery approaches differ. Tier 2 responds well to structured recovery practices.

Will taking a two-week vacation fix Tier 2 fatigue?

A vacation helps temporarily, but without changes to how you work with AI tools, the fatigue returns within weeks of being back. The issue is structural — the habits and systems that caused the fatigue are still in place. Tier 2 requires practice changes, not just rest.

I’m a senior engineer and I’m embarrassed to admit I’m struggling with this.

You are not alone. Tier 2 is extremely common among senior engineers — particularly those who care deeply about craft, who have the most to lose from skill erosion, and who are often the most reluctant to admit vulnerability. The identity dimension of Tier 2 is especially sharp for experienced engineers.

How is Tier 2 different from Tier 1 fatigue?

Tier 1 is awareness without functional change — you notice something is off but your work and identity are still intact. Tier 2 is functional change: your relationship with code has measurably shifted. You might not trust your own output, feel distant from your work, or notice that problems you used to solve easily now feel opaque.

Can I keep my job while recovering from Tier 2 fatigue?

In most cases, yes. Tier 2 recovery is about changing your practices, not quitting your job. The key is finding sustainable boundaries with AI tools while communicating honestly — without calling it “AI fatigue” — with your manager about workload and quality expectations.

How do I know if my fatigue is getting worse vs. getting better?

Watch for these Tier 2 escalation signals: you stop reading code you did not write with AI, you cannot explain your own recent PRs in a 1:1, you find yourself avoiding implementation tasks, or you feel nothing when something you built works. If several of these are true, your fatigue is deepening — act now.

Continue exploring

Recovery