Post

"LLMs can help non-programmers with no technical background get a foothold in doing something that used to require a LOT of technical knowledge to get started in, putting them in a better position to learn the technical stuff over time by providing a less steep learning curve" feels uncontroversial?

As a former academic I take some issue with this, since just having the tool do work for you is not learning. It's going to a restaurant and pretending you're a chef. If you use the tool to help understand what's going wrong with what you made, i.e. as a tutor, that's a lot better

So the situation I was really talking about is one where you have to learn a lot of technical skills before you can get to the thing you want to actually learn.

In particular, if you want to make videogames, you generally have to learn a lot about programming before you can really get to the part where you're designing games. My hope is that LLMs will facilitate people being able to get to a place where they can design a game, test it, and iterate on it...

This sort of sneaks in the assumption that there's actual design work involved. In my experience, the kind of person who actually puts care and thought into their game design is the same kind of person who doesn't usuallyhave much trouble learning the technical side in the first place.

Clearly you don't have very much experience with the wide variety of kinds of people who are interested in making games and put care and thought into their game design.

Your subseqent comments make me think that your primary interest is your own success, not in more people learning to make games. There isn't really anything wrong with that, but those two things are fundamentally at odds. More people making games means more competition for eyeballs. Simple as.

This is not to say that slop is good and that we shouldn't have mechanisms to ensure that better games get focus. However, slop is an inevitable consequence of broader access to the capability to produce things. Before AI slop it was Unity asset slop.

Any change that makes it easier for someone who is inexperienced to start building also makes it easier to produce slop. Most people are bad at most things, and many of them do not actually want to get good. Of course, many more GOOD games will also be made, but Sturgeon's Law as a constant,

I think accusing me "only thinking about my own success" is an incredibly bad-faith read of my argument and it's making me reconsider whether I really want to engage with you further. What I'm thinking about first and foremost is _discoverability_. I like good games. I want to play and find them.

More good games may get made in absolute numbers, yes, but the mere fact that the effort and thought that goes into a good game is not something that's impacted much by the use of LLMs means that slop is _much_ faster to push out in much greater volumes than gold.