Gamma.app Made Me Faster. I'm Not Sure That's Only a Good Thing.

This post was written while listening to this beautiful song by Nils Frahm.

I have been making presentations faster than I ever have before. Gamma.app is one of those AI tools that turns a conversation into a structured deck almost instantly. You describe what you need, the slides come together, the layout holds, the logic flows. I was impressed. I still am. I will use it again.

But somewhere in the middle of one of those sessions, something shifted. I was adjusting a slide, moving quickly, and I remembered sitting with a presentation for a long time, going back and forth on a single heading, getting the typography exactly right, building something slide by slide with full attention on every decision. I remembered the feeling of a finished deck that looked the way it did because someone had cared about every pixel of it.

I do not get that feeling from a Gamma session. I notice its absence. I cannot tell yet whether that matters.

The Speed Problem

I have spent a long time in technology. I have seen enough crises and enough urgency to know what speed actually delivers. The answer, consistently, is less than you think. The work that held up, the decisions that turned out right, the things that lasted, almost always came from the slower process. The one where someone was stubborn enough to go back and check, to rethink the assumption, to sit with the discomfort of not knowing yet.

AI's gift is speed. That is also, for me, its most unsettling quality.

The core of the problem is this: the output appears faster than the understanding does. Gamma gives you a presentation in minutes. Claude builds something you could not have built alone because it knows things you only half know. The output arrives before you have fully earned it, and I find myself in the strange position of producing things I am proud of while being uncertain about what I actually learned in the process.

There is a specific moment I keep returning to. I was building something with Claude recently, a section that required understanding I do not fully have yet. Claude filled it in. The result was correct. But I knew, while reading it, that I had not gone through the thing that would have made it mine. The understanding was borrowed. The output was real.

What I Am Actually Worried About

I do not think AI is making us stupid. That is not the right frame.

The concern is more specific. It is about the relationship between struggle and competence. Something happens when you have to figure something out the hard way, when you are forced to sit with a problem long enough that it reveals its structure to you. That process is slow and uncomfortable and it builds something in you that the output alone does not capture.

When AI shortens that process dramatically, the output is there. The depth behind it is sometimes borrowed. The question I keep asking is what accumulates over time when most of the hard parts are being done by something else.

I keep building frameworks to manage this. Task lists with enough detail to slow me down deliberately. Constraints I set before opening a tool. I am not very good at maintaining them. The speed of the tools is more seductive than good intentions can resist consistently. Dopamine is speed. I wrote about the rise of dopamine culture before, and the logic applies here as directly as anywhere. Fast action, fast result, fast feedback. The loop is tight and it feels productive. What it produces over time is the question.

What I Also Know

I should be honest in the other direction.

There were slow, careful, detail-obsessed work environments that were also full of ego, politics, and performative thoroughness that had more to do with status than quality. The meetings that lasted all day to produce what could have been an email. The revision cycles driven by whoever had the most seniority rather than by what was actually better. I do not miss those things. I do not miss the human friction that had nothing to do with making the work good.

What I miss is more specific. The version of slow work that was genuinely about the work. A keynote built with care, white background, clean screenshots, good typography, every slide a considered decision. The kind of thing you could look at and feel the attention that went into it before reading a single word. That care was its own argument.

I still make things that way sometimes. I think I always will. Not out of nostalgia, but because certain kinds of work require a quality of attention that speed is structurally incompatible with.

Where This Leaves Me

I am not at the point of revolt. The tools give me something real: the ability to build things I could not have built alone, to move in directions that would otherwise be closed to me, to keep producing when the alternative might be producing nothing. That matters.

But I am watching the speed carefully. Not with hostility. With the kind of attention you give to something that has genuine power over you and that you want to understand clearly before it changes you in ways you did not choose.

The days ahead will show what this adds up to. In the meantime, I am still taking care with the typography.

---

Related reading:

• The Polished Lie: Why Great AI Execution Still Feels Wrong -- on the difference between output that is technically perfect and work that actually belongs to you. The same tension, from the angle of music and writing.
• The Rise of Dopamine Culture -- the structural case for why speed feels so good and what we lose when we stop sitting with things long enough to understand them.
• Claude Code Changed How I Think About Marketing -- what happens when a growth marketer starts touching the stack directly, and what the gap between output and understanding looks like in practice.