If I had to sum up the state of AI in 2024 in a single anecdote, it would be this: Once upon a time, I followed a concept artist on X.com. This individual worked in the gaming industry for years and I thought their work was outstanding. Around December 2024 they posted something about how people are relying on AI-generated images instead of artists, that artists are being replaced, and that AI is training on artwork without permission, and all of it was horrible to the industry.
I responded with a comment sharing my take, which was, I ultimately don't think it matters, because even if people use AI artwork, it'll always be a cheap counterfeit compared to the real thing. I was in the middle of writing a follow up post to defend his craft, when I suddenly couldn't see the original tweet anymore. It turns out this person blocked me for my comment.
The point I attempted to make was simple: if I want an artist, I'll hire one. There's a reason I don't hang up AI artwork around my house. Some people do, but they probably weren't buying $1000 originals from artists before the AI image gen explosion. The expense is in the craft, and the thoughtfulness of every stroke that goes into the final piece.
The question I have rooting around in the back of my head is, will that change?
While building Sloper in this environment, with the tools available on the market today, I think designers do have a leg up on engineers and other AI-enthusiasts. When everyone can build, the craft matters more than ever, and there's no question that these current AI app builders and tools output designs with an obvious taste gap. The output looks right but usually doesn't hold up under scrutiny. The solution may be present, but the UX was likely flubbed. It's still a great starting point, but this is where the designer shines.
Designers can design every corner of the product and build it to their detail oriented specifications. The gap between a design and the live product is a matter of prompt quality, and that can always be refined. In the current environment, it feels like the Product Designer has the superpower, for now.
I think product designers have a 9-12 month runway before AI's app layout output gets good enough that the designer can be bypassed altogether. I say this because of how far these image and video models have come since Will Smith Eating Spaghetti. It's only a matter of time before models have trained on enough designs to understand component hierarchy and architectural reasoning in a way current models don't. The outputs will continue to improve and designers will slowly become less necessary.
A few thoughts on the road ahead:
- Outputs will be incredibly uniform and uninspiring, meeting a threshold of acceptable
- More people will build out their own ideas than ever before
- No one has to be left sitting on the sidelines, anyone can build anything
- Standing out will still require a human touch
It's an interesting crossroads to stand at, to have worked at developing a skillset for years, and then watching AI learn it at a pace 1000x faster. All the while knowing that it's never going to stop or slow down. I think it ultimately boils down to two mindsets: This is going to replace me and make me obsolete, or I've never been more capable to pursue any dream I can imagine.
I stand firmly in the camp of the latter.