The issue isn’t the quality or the visual result. It’s not just slop because it looks bad.
Art is as much about the journey as it is about the destination. It’s the friction of someone going through each detail, making a choice. It’s the artist taking time to figure out how they want to convey what they’re feeling and thinking.
The reduction of this friction culminates in the reduction of the piece’s essence.
Yeah I understand your point because I've made music and design for 25 years the old "manual way", and there is definitely less friction with AI than there is the manual way.
But what interests me about AI is that it is possible to do things that you couldn't do before particularly in processing images. For these images I took photos of the bottles myself and used other images and the AI can place the camera anywhere in a scene and place objects anywhere in the scene, and things like that. You can create almost any image you can imagine with the right sources and the right prompt, and even here there is some friction because many images do end up looking like slop and you have to prompt it right to get something non-slop like out of it (though not always).
This ability to manipulate and process information is what makes AI interesting to me, although I'm not sure I adequately explained it. It feels like a new tool like Photoshop except across a different dimension. I feel like we've barely scratched the surface of what's possible to create.
> You could potentially ask an LLM to explain it, but I would venture a guess that wouldn't feel like the right source to help you.
I don’t think this works, at least for me. Even on the 5.5 versions of Claude it’s still arduous to read at this point.
> You could use the LLM to help you identify source material written by humans in order to learn anything new. It doesn't have to "do all the work for you".
This probably depends on the shop, but humans aren’t writing docs anymore. That was the first thing to go, sadly.
I hear what you're saying. I often tweak the CLAUDE.md file to improve the quality of its output. It's an ongoing experiment.
As for "humans writing the docs", you raise a good point although it's not quite what I imagined when I wrote that line. I was thinking more about having the LLM/Agent create a list of verified output by humans with reputations as experts in whatever field interests you. This is important when pushing the boundaries of research, although it's overkill for wanting to know whether or not a new technology could be useful in your existing project (or the next one).
What use case did you have in mind when you brought up the docs situation? Sounds like you had some direct experience with that one.
I think a distinction is clearly emerging between docs intended for humans, and docs intended for agents. The latter is where all the Markdown-ridden slop should go; I'm not insulting humans with that.
Why not just apologize and have Claude rewrite any code that's similar? Whether you did it intentionally or not, you probably agree that you can't keep using his code and that it would pretty bad if someone did something similar to you.
Sometimes the sense of satisfaction in creating a piece of art is that it exists at all as an identifiable representation of the original intention. I don't buy these purity tests that disqualify art based on arbitrary measures of craft. That went down the urinal with Duchamp's Fountain.
That is orthogonal to the issue of game developers exploiting technology to disenfranchise labor and produce a shoddy product. There certainly should be economic breathing room in the richest societies on the planet in the wealthiest era of human history for artists to spend time crafting work how they like.
Many artworks (across mediums) have parts where the artist isn't directing their creative decision-making so much as reaching for the nearest stock asset / procgen tool, or googling "leather texture photoshop trick" instead of carefully studying the usage-appropriate types of leather and looking up authentic period nubuck references and hand-painting it to get it just right, etc. Some "art" assets really are utilitarian and fungible subunits in the context of the full artwork they're used in, just like the code one would LLM. And many games are sprawling works, with that much more room for such.
Going whole-hog on artistic intentionality for every little detail is just bikeshedding.
Strong evidence for this: the cottage industry of reusable gadgets in every creative field. Presets and sample packs in music[0], brushes and textures for illustration, models and rigs and so on for 3D. You could go on forever listing it.
There are lots of artists using generative ai in the same way there are artists using photoshop. It’s about control and intent, and it only gets weird with gen ai because you can give away both and still get results. Many people do take the role of a non creative manager, but that doesn’t meant you have to.
The difference is that game code isn't something that the player sees or needs to care about. It's possible to write two functions that look different but have identical properties for anything that ends up being perceived by the player. For assets, the visual difference is directly exposed to the player. A sixth finger on a hand is an issue in the way that as line of dead code isn't.
Adding to the fact that people who write code are in general much better compensated in most industries than people who produce art, having software replace artists feels like "punching down" in a way that having software replace people who write software doesn't. As a software engineer, Ilif I use AI to write a bunch of code for myself, I'm much better able to understand the tradeoffs to my future employment in a much more visceral way than I ever could for what using AI-generated art might mean for an artist. Even with all of the recent proclamations that code is "solved" and fearful predictions that programming jobs are going to stop being so cushy, the job market for the average software engineer is probably at least an order of magnitude better than it is for the average artist in terms of compensation and job security, and AI or not, I don't see that changing any time soon.
That's not true though. Bad code will manifest in a lot of ways. Affecting players and how they perceive the game.
See Cyberpunk 2077, the art is pretty good, but the game is extremely buggy. I stopped playing because of that
I agree. It does make me think these tools need an irrevocable signature on the files, if that’s even possible. You can ask someone you’re commissioning if they use AI, and they can just lie.
Now should have Jagex reviewed it? Of course. However, there’s just this mistrust in art now. I’ve commissioned two things from artists. I asked both, as for my work I want to fund actual artists. However as these tools get better, I’m concerned that trust won’t be good enough.
Fake edit: “Well if it’s so good you can’t tell the difference then why’s it matter” completely misses the point so let’s not bother with that.
There is probably a marvelous software core in many large enterprises but at least in my experience it’s surrounded by layers and layers of very basic stuff. Read from a database, write to a database (often not even with any logic dealing with parallel read/writes). Excel Macros. Basic forms to submit data. Scripts to print some stuff from a database. …
I’ve been experimenting with this idea that I can implement features faster and more accurately than the big model + agent swarm paradigm. I also have been trying this because of what the post brings up when it comes to AI burnout.
For me, writing the code (literally typing), shifting files around, renaming things is slow. So I’ve been using small models + much stronger grasp on the reins. Similar to what this post mentions. It’s been great, the decisions come from me. I tell it to make a domain class with these fields and these invariants. It does it in a split second, I look it over, tweak it and move to the next part of implementation.
When I experimented with swarms it would spend an hour just having agents adversarially review to decide some pretty trivial details. This way, when someone asks me a question during review, I can answer it. When I need to dive back in, I know what to look for and where to look for it. I have way more connection to my work, than a few months prior.
It's the difference between constantly reviewing someone else's code and getting to write the code yourself.
There's both deep satisfaction and deep value in having an understanding of the past, present, and notional future of the system (especially at scale). If I use agentic tools to do the actual coding, I never feel like I have as deep an understanding of the implementation, it's inherent assumptions or lessons learned. Over time, my understanding of the system atrophies badly enough that it becomes difficult to steer the LLM, instead of the other way around.
This is what makes those "unicorn"/great managers who both manage well AND maintain a clear understanding of the system over time so amazing.
Early copilot was exactly what I wanted for a long time. A smarter autocomplete. Then they started taking it too far with just filling in the entire file.
reply