It's not ai slop, it's low-input

I want to suggest we start using the term, “low-input” instead of slop, because by doing so, you can find a path to better outputs.

Let’s be clear, AI slop is generated at the request of humans. Yes, a highly capable AI model does the work to create the output, but a human asked it for an output.

We call it slop because it all tends towards mediocrity at an unprecedented scale.

Calling work slop is unhelpful because it’s just a complaint without a solution. We could throw the focus back on people and call it lazy, but that’s rude and doesn’t help anyone.

Low-input is less judgemental and suggests the solution: higher input. It’s not catchy, so it won’t catch on, but that has never stopped me before.

If we think about the reason a highly capable machine churns out mediocrity it’s because most humans don’t possess the visual design language or literary style pointers to actually prompt and then iterate on an output to get what they want. I could fall into the trap of saying most people don’t know what they want, but that’d be sloppy wouldn’t it. Clients always know what they want, we just have to listen in the right way.

I solve problems by building digital tools. In this instance I’m creating a cheat sheet that you can use to look up different visual styles, learn some visual design language and see example, detailed prompts to better prompt your way to less “samey” outcomes.

Stay tuned.