So, I think this article is circling something a lot more foundational when it comes to people’s disdain for AI. If you put more thought, more effort, and more originality into your work, you get something that starts to feel like an actual artistic expression. If you just type in a basic prompt and accept whatever comes back, that shows too. And I think that is a large part of why people feel insulted by this stuff in the first place.

It’s hard to respect a piece of work or take it seriously when it feels like the person who made it didn’t think it was worth giving much consideration. You’re asking someone to stop, look, and care about your poster and whatever it represents, but you haven’t given them much reason to believe you cared about it yourself. If you can’t take it seriously, how should we? It isn’t really about whether it took five minutes. It’s about whether there was any thought behind those five minutes.

So, I read this article as if it had the following title: “If you insist on using an LLM to generate visuals, you still have to put some thought into them.” But then, that starts to defeat the point for anyone who was using an LLM to avoid that effort in the first place, doesn’t it? You can make the process easier, but you can’t just assume that having produced something means you’ve made something worth sharing.

At the end of the day, I think there are valid uses for LLMs and there are uses that don’t add much of anything. As a consumer of content, the ones I respond to tend to be the ones where the LLM’s involvement isn’t the most obvious thing about the work. I want to see the person’s idea come through. A very light touch, not a deluge. This can absolutely apply to creating visuals, but it requires judgment, revision, and a willingness to do more than accept the first thing you’re given.

And it seems like a lot of people aren’t willing to put in that effort. That’s fine. Their work will continue to be distinguished by that lack of care, while people who are willing to put more into it will stand out. The initial novelty only carries something so far. Once you start picking up on the familiar patterns, that sense of something being slightly off becomes hard to ignore.

So, moving forward, I’m not so sure human creativity and input are things that can simply be outdone and made irrelevant. Maybe they’re an additional layer that keeps giving us somewhere further to go. An LLM can imitate what people have already brought to the table, but given access to that same LLM, we can bring our own experiences, ideas, and judgment into the mix. The tool gives us a starting point. I don’t see why we should assume it also determines how far we can take it.