Stop treating depth like a personality trait Link to heading

If shallowfication is produced by incentives and interfaces, you do not solve it with better vibes.

You solve it the way you solve any persistent systems failure: you add constraints, you change defaults, you introduce checks, and you redesign the feedback loops that keep the failure alive.

Personal habits matter, but they are not the whole story. Telling individuals to ‘just focus’ while surrounding them with interruption-by-default products is like asking a pilot to fly through instrument noise and blaming them for missing the signal.

So here’s a more honest approach: treat depth like a requirement. Something you design for, protect, and verify. Not something you hope emerges.

Layer 1: personal overrides Link to heading

Start with what you can control. Not because it is sufficient, but because it is available today.

The move is simple: stop letting every app negotiate directly with your attention. Put a gate in front of it.

If you want something concrete, pick one high-value block per day and protect it like an appointment. Phone out of reach. Notifications off. One task. When you drift, you return. No drama. This is not a productivity hack. It is reclaiming a basic capability.

Then add two small pieces of friction that feel almost stupid, which is how you know they work.

First, delay consumption. When you feel the urge to check something, wait thirty seconds. You are not proving a moral point. You are breaking the reflex loop long enough to choose.

Second, finish things. Read the article you opened or close it. Watch the video you chose or stop. Shallowfication thrives on half-starts. Depth is often just completion, done repeatedly.

These are local fixes. They reduce damage. They do not change the upstream factory. For that you need product-level changes.

Layer 2: product defaults and stopping points Link to heading

If you build products, you have leverage. Use it.

Most teams already know how to increase engagement. They have playbooks, dashboards, and a catalog of patterns that remove stopping points. The design question is whether you can build a product that respects the user’s goals even when those goals conflict with the product’s short-term metrics.

That requires defaults that make depth easier, not harder.

Start with stopping points. Infinite scroll, autoplay, and algorithmic feeds are not neutral. They are continuation machines. If you insist on using them, then add explicit exits that do not feel like failure. Natural ends. “You’re caught up.” A pause screen that does not guilt-trip. A prompt that asks what the user intended to do before the loop started.

Then make time visible. Not as shame. As instrumentation. If a user cannot tell where their time went, the system is operating without a dashboard. That is convenient for the product, not for the person.

Next, demote interruption. Notifications should be rare, meaningful, and user-shaped. Many products treat notifications like free distribution. They are not free. They are a tax on attention and a steady source of cognitive fragmentation.

If you want a clean heuristic: if a notification would be annoying from a human, it will be annoying from an app. Your users tolerate it because they are polite and addicted, not because it is good design.

Finally, stop using social pressure as a retention strategy. Streaks, badge loops, and public counters can be useful in narrow contexts. They become corrosive when they replace intrinsic motivation with fear of breaking the chain. If the product needs a guilt mechanism to stay sticky, it is not helping the user build a habit. It is managing a dependency.

None of these moves is exotic. The hard part is choosing to accept slower growth in exchange for a healthier relationship with the user. That is a business decision, which brings us to the third layer.

Layer 3: institutions and incentives Link to heading

Shallowfication persists because the costs are distributed and the rewards are concentrated.

This is why “ethical design” speeches do not move markets on their own. The system responds to what gets rewarded.

So we should expect a mix of pressure sources: consumer demand, procurement standards, regulation, and professional norms. Not a single heroic fix.

At a minimum, we need stronger expectations around transparency and control. Users should be able to understand why something is being shown to them, and they should have meaningful ways to change that behavior. Not buried toggles. Real control.

Education matters too, but not as a poster on the wall. Attention literacy should be treated like basic civic infrastructure. People need a working model of how feeds shape beliefs, how proxies distort behavior, and how to slow down long enough to verify. The goal is not to make everyone a philosopher. The goal is to keep people from being easily steered by default.

What “designing for depth” looks like in practice Link to heading

Here is the practical stance I am arguing for:

Build systems that assume users will sometimes want less stimulation, fewer interruptions, and more time to think. Then make that path easy to take.

That means products should support intentional modes, not just maximal engagement. Reading modes that actually reduce distraction. Social modes that reward conversation over performance. Work modes that protect deep time instead of slicing it into fragments.

It also means admitting a quiet truth: not all engagement is good engagement. If your metrics cannot tell the difference, your product will not be able to either.

The missing lever is user-owned filtering Link to heading

Even with better defaults, we still have a structural problem. Most people do not control the filters that shape what they see. Platforms do.

That becomes unacceptable as AI gets better at persuasion, personalization, and behavioral prediction. When meaning-making is mediated through systems optimized against depth, “choice” becomes a polite story we tell ourselves.

So we need a stronger lever: portable, user-owned preferences that can shape feeds across products. A way to express values and boundaries in a form systems can respect, and users can carry with them.

That is the bridge to the next part. If we want real agency, we need more than tips and nicer settings. We need a new layer in the stack.