Is it Slop or is it Craft
Friction as a Feature by Emil Kowalski (great piece as usual) made me think. What happens to craft when the barrier to entry is removed. How can we make products stand out in an era of fast-fashion software.
Life becomes great again when you realize this has all happened before and it’s just the circle of life. New tech comes in and makes the old ways obsolete, creating abundance in the process.
Going all the way back to the early 1500s, the printing press caused the original ‘content slop’ panic. It let more ideas and voices be published, while the scholars complained about worthless material filling the world.
Digital photography brought the cost of another attempt close to zero. It did not kill professional photography, it moved the selection from before production to after production.
A more recent example is website builders and templates. They allow non-specialists produce work that once required a professional team before. Tools like Webflow and Framer let designers ship websites without writing a line of code.
It is true that these tools made many outcomes feel less special and distinctive. At the same time, recognizable excellence still happens beyond the baseline.
Now, do we actually want to remove friction? Let’s look at these three neat concepts:
Generative friction - producing the first draft. AI is excellent at it and let’s be honest, we always wanted this. Think countless templates, libraries and boilerplates.
Formative friction - developing judgment while building. Struggling with constraints, looking for alternatives, learning systems, receiving critique, revising and crafting the details. This is something AI can assist us with, and something we want to preserve. Use AI to learn, not bypass it.
Shipping friction - putting something into the world. Making something actually useful, understanding users, tradeoffs, edge cases, maintaining it. Tools can help, but the responsibility cannot be outsourced.
The old world of friction and gatekeeping did not guarantee craft. Just remember bad enterprise software, broken websites, or expensive products protected not by excellence but distribution. Lowering barriers admits people with strong ideas who previously lacked money, credentials, time, or geography.
My bet is that AI is not killing craft.
It removes the friction of producing an artifact. It makes it easier to avoid obviously bad work, but it does not automatically make it unusually good. It moves the bottleneck from execution to judgment, taste, and original ideas.
It lets creators and builders spend less time operating tools and rebuilding defaults, and more time crafting the details and working on what matters.
Abundance is noisy, but it makes craft, curation, taste, trust, and distribution even more valuable.
Net positive imo.
The formative friction insight is really about measurement systems. What you're describing as "judgment, taste, and original ideas" is actually the ability to measure "which outputs are worth shipping versus which are just generated volume."
When execution friction was high, judgment came before production - you made a deliberate decision before you spent a week building. Now that execution is cheap, judgment has to come after production, which flips the problem completely. The scarcity isn't execution anymore. It's the measurement system for "does this version deserve to go into the world?"
That's formative friction. Not "can I make a first draft" - anyone with an AI can do that. It's "can I look at 12 versions and articulate why version 7 is right for this specific person instead of the generic pattern."
The confusing part is that this looks like taste or opinion when it's actually measurement. The photographer who shoots 400 frames and selects one isn't being artistic - they're measuring "which frame contains the moment I intended to capture, not what the camera defaulted to." The writer who generates 12 openings and keeps one isn't being picky - they're measuring "does this version make the reader feel what I intended, or does it just sound like good writing."
Once you frame it that way, craft becomes reproducible. Judgment becomes learnable. It's a measurement system you can teach: "here's how to measure whether an output actually solves your specific problem versus just being technically correct."
The photography comparison is the right lens. Digital didn't kill craft, it moved the selection boundary — from "can I afford a shot" to "which of 400 shots do I keep."
AI writing is doing the same thing. The selection pressure doesn't disappear, it moves. Before: "can I generate this?" After: "which of the 12 versions is actually worth saying?"
The irony is that craft might matter more now, not less. When a great photographer worked with film, most people couldn't tell whether a bad photo was a bad idea or just bad technique. Now both are fully separated. Same with writing: a weak idea expressed perfectly is obviously a weak idea. The AI removes the technique excuse.
What changes is where the judgment gets applied. Building tools that help people apply judgment better — at the selection and curation stage — feels like the more durable angle than tools that just add more volume.
Your distinction between generative and formative friction is the useful part. An AI first draft can get you past the blank page, but it can also smooth out the odd phrasing that made the idea sound like you before you've had a chance to judge it. That's why I built DictaFlow. Speaking gets the rough draft out quickly and keeps the author's actual words. AI only cleans up transcription noise. Fast creation helps. Handing over your voice doesn't.
The formative friction point is the one I keep coming back to. Removing the work of producing a first draft is obviously useful, but removing too much of the struggle around judgment can make the output faster without making it better.
Good framing on formative friction — that's the layer AI should assist with, not remove. Building an AI-assisted reply tool for Threads creators showed me the split is clean: drafting is now cheap, but knowing which conversations deserve your time is still pure judgment. The tool removes typing friction, not taste. Net positive, agreed.