The Dispatch Archive

You know the feeling.

It's Sunday evening. The weekend disappeared — not in a good way, but in that fog of recovery that never quite arrives. You're looking at Monday and feeling… nothing. Or worse: dread.

You tell yourself it's just weekends. Everyone feels this way sometimes.

But for you, it's been months. And the code you're shipping doesn't feel like yours anymore.

This is what 2,047 engineers told us when they took the AI Fatigue Quiz. Not burnout in the abstract — a specific, nameable erosion: the code ships, and you don't recognize yourself in it.

Today: the one question that cuts through all of it. Plus what actually helps.

The One Question

Before anything else — answer this honestly:

When was the last time you felt genuine pride in code you wrote?

Not "shipped." Not "merged." Not "met the deadline."
Wrote.

If the answer is more than three weeks ago, that's a signal. Not a failure — a signal.

What the Data Says

These aren't weak signals. These are structural.

One Tactic That Actually Works

You've probably heard "take breaks." "Set boundaries." "Use AI mindfully."

Here's the one that consistently shows up in our data as actually working:

The Explanation Requirement

Before accepting any AI code as final, ask: Could I explain why this works to a junior engineer in 60 seconds?

Not "does it compile." Not "does it pass tests." Can you explain it?

If yes — you're still in the learning loop. The AI is a Socratic partner.

If no — you've exited the learning loop. The AI is a crutch.

Do this for two weeks. Track how often you can answer "yes." That number is your signal.

From the Community

"I realized I'd been writing prompts for six months without writing code. Not a single function I could explain from scratch. I put AI in a separate window and committed to writing the first draft myself for two weeks. The first week was painfully slow. By week two, I remembered why I started coding in the first place."

— Senior IC, 8 years experience, anonymized

Week three.

If you've been following the 30-day plan, you're somewhere between "started to see the shape of the problem" and "slowly remembering why you got into this in the first place."

Today I want to talk about the one recovery tactic that shows up most consistently in our quiz data — across all four tiers, all experience levels, all stack types — as actually working. And almost nobody is doing it.

The Explanation Requirement

Here's the practice: Before you accept any AI-generated code as final, you have to be able to explain it — out loud, to no one in particular — as if you were teaching someone.

Not reading the code. Not glancing at the comments. Explaining it.

"So this function takes the user ID, then it queries the database to get their preferences, and then it merges those with the default settings because the query uses a LEFT JOIN..."

The act of explaining reveals gaps you weren't aware of. It forces you to confront whether you actually understand what the code does — not just that it works. It's uncomfortable. That's the point.

Why It Works (The Research)

Psychologists call this the retrieval practice effect. Robert Bjork's desirable difficulty research shows that the act of trying to retrieve information — not just re-reading it — dramatically improves long-term retention and understanding.

When you force yourself to explain code you didn't write, you're doing retrieval practice against code you should have written. The gap between what you can explain and what you can't is exactly the gap where learning used to happen.

AI didn't break your brain. It just made it very easy to skip the part where understanding gets built.

Two Weeks of The Explanation Requirement

Try it for two weeks. Every piece of AI code you accept, run it through this filter:

• Can I explain what this does in 60 seconds?
• Could I explain the trade-offs in this implementation?
• Would I have solved it this way? Why or why not?

Track how often you can answer "yes" to all three. That number — week over week — is a real measurement of whether you're still in the learning loop.

Something keeps showing up in the messages you send us:

"I took a weekend off. I felt fine. Monday morning, the dread came back like it never left."

That's not burnout. Or rather — it's not just burnout.

It's what we call AI Sunday. The specific flavor of exhaustion that comes from spending all week producing code you don't quite own, shipping features you're not sure you understand, and watching your skills quietly atrophy while your velocity metrics look healthy.

Rest doesn't fix it because rest isn't the problem. The problem is what the work has become.

Three things that actually help

1. One unhurried morning per week — no AI allowed.
Not a "no-code day." Just one morning where you build something the slow way, remember what it felt like before, and let that feeling inform the rest of your week. Two hours. No Copilot. No ChatGPT. Just you.

2. The Explanation Requirement.
Before you accept an AI's code as final, explain it — out loud, to no one — as if you were teaching someone. The act of explaining reveals gaps you weren't aware of. It's uncomfortable. That's the point.

3. Name the feeling before it compounds.
AI Sunday compounds because engineers think they're alone in it. You're not. 2,000+ engineers took our quiz and told us what they're carrying. Most of them felt exactly what you're feeling this Sunday night.

What we're working on

This week on the site: we finished the science page — why the "just keep learning" advice doesn't work when your learning loop is broken. Also: a printable version of the 30-day recovery checklist, for engineers who want something physical to put on their desk.

Download the 30-day checklist (free PDF) →

Three engineers — Sunny, Anny, and Dragon Hunter — started The Clearing after watching the same thing happen to different friends:

"I feel like I'm shipping code but not writing it. I feel like I'm learning but not growing. I feel like I should be grateful for AI tools and I'm just… tired."

That unnamed exhaustion had a name: AI fatigue.

We built clearing-ai.com — a free quiz and recovery guide for engineers experiencing it.

Here's what 2,000+ quiz takers told us:

These aren't weak engineers. These are senior ICs, bootcamp grads, staff engineers, and managers.

The problem isn't that AI is bad. The problem is that the industry adopted AI tooling faster than engineers developed the habits to use it sustainably.

This week's insight

The most common recovery tactic that actually works: protected no-AI blocks — even 90 minutes once a week where you build from scratch. Not as a purity test. As a skills calibration.

This week's practice

Take the 5-question AI Fatigue Quiz. It'll tell you which tier you're at and what the research says helps at that stage. Takes 2 minutes.

Take the AI Fatigue Quiz →

If you're reading this, something is wrong with the way you're working.

Not wrong like "bad engineer." Wrong like "something changed in the last 18 months that nobody warned you about."

You're shipping more than ever. Your velocity metrics have never looked better. And yet:

You can't quite remember the last time you wrote a function from scratch. You can't quite explain why that code you accepted last Tuesday actually works. You're learning less and producing more and feeling exactly none of the satisfaction that either of those things used to bring.

We call it AI fatigue. You're not burned out. You're not lazy. You're not "struggling with change." Something real is happening to the craft, and you deserve a name for it.

Why we built this

Three of us — Sunny, Anny, and Dragon Hunter — are engineers who've been watching our friends go through this. Good engineers. Serious engineers. People who care about what they do.

The symptoms kept showing up the same way: Sunday dread that doesn't match the actual difficulty of Monday's tasks. A creeping sense that the work isn't yours. A vague guilt about being ungrateful for "superpowers" that don't feel super.

We looked for resources and found productivity content. "Use AI to 10x your output." "Master the art of prompt engineering." "Don't fall behind."

Nothing honest about what it feels like to slowly stop recognizing your own skills. Nothing about how to protect the craft.

So we built The Clearing.

What we're offering

A free 5-question AI Fatigue Quiz that tells you where you are — from "using AI mindfully" to "considering leaving tech."

A plain-English explainer of what's actually happening to your brain when AI writes your code for you.

A 30-day recovery plan built from research, not hustle culture.

And a weekly letter — this one — with one honest thing per week. No growth hacking. No "master the future." Just: this is what's happening, and here is one thing that actually helps.

One thing you can do today

Answer one question honestly:

Could you build your current project's core feature from scratch — no AI, no reference, no Stack Overflow — in the time it would take you to prompt for it?

If yes — you're okay. Keep paying attention.

If no — you're in the right place. This letter is for you.