This got shared on Mastodon, I think, but I don't know by whom. It got my mind all in knots about creativity and my opinions on AI, so I tried to write some of it down. I think I got most of it out!

The problem with LLMs democratising creativity is they're essentially creating an average. It's not great things, it's not bad things, it's just things. The more you consume of things created in this way, the duller your expectations become. When people only look at pictures that please them, or read things that are easy, their ability to rationalise and process conflict is diminished. I feel like it started with the endless scrolling of Instagram Explore, which TikTok then turbo-charged (I'm not an historian - it probably started much earlier. Probably with the printing press). When everything is good, "good" loses all meaning. Pluribus conjures a lovely image of this. You can't just have the middle of the road; the human experience is rooted in conflict. A happy ending is redundant when the beginning and middle are also happy. Even Christmas movies have conflict.

There are benefits to having the aggregation of all human knowledge at your fingertips¹, but you're delusional if you think it makes you Picasso because it doesn't. It reminds me of some people's attitudes to modern art: "that's so simple - even I could do it!", but in art, the execution is only half of it (maybe more, maybe less). Knowing what to draw, paint, write, play is frequently the hard part. With a concept, you can muddle through any old execution and say "here's my picture of a tree" that looks nothing like a tree. The art is in your concept and your execution and the fact that you did it, not that it was done. The struggle in the execution breeds refinement of future concepts, which then pushes you to practice and improve the execution. Spending a week toiling at a shallow concept encourages you to make better concepts, but if you can fart out any old concept and have an LLM make you something in ten seconds, you never learn that your concept was fundamentally lacking. Even if you end up with a pleasing end result.

There's a lot of people writing about blogging at the moment. For some reason, blogging's a new year's resolution, which I love, but you have to start by doing it for the joy of doing it. If you start a creative endeavour for fame or money, you're doing it for the wrong reasons. Those are sometimes nice by-products (or so I imagine: I've experienced neither fame, nor profit from blogging, but others have!), but it's supposed to be fun. Even if it isn't all fun (I have been taking photos since I was 17 years old, and I still hate processing them. I find it so tedious), if you don't enjoy it more than you hate it, I think you'll struggle to do it enough to get good at it. And once you're OK with that, you only have to care about one person's opinion on whether what you're doing is good: your own. When you externalise validation of your creativity, you give away control of it. One of the reasons I left Instagram was that my selfies got more likes than photos that I was genuinely happy with. What's the point in trying to make something good when I can just shove a camera in my own face and get validation? There is none, and that's all these platforms offer when it comes to your creativity being personally fulfilling: maybe some other people will like what you do, and then that means it was worth doing.

And that's why I think blogging is a thing everyone should do. Make yourself a little island, and write just to write. You can write to document something specific, or just a glorified journal. You don't really even have to write. You don't have to make something good; you're in very little danger of having an audience when you first start out anyway. Just do something. Make yourself do it, even when you don't feel like it; it'll hopefully make you feel good to have two little sentences or a photo or a drawing or 15 seconds of music that wouldn't have been there otherwise. Sometimes I think the times you grow most are when you do things even when you didn't want to; it's purposeful practice, and you earned the growth.

But do it yourself, and for yourself, on a platform you own if you can. It doesn't need a pretty design, or some complex coding; I just think it's important that it's yours. There's a simple but pure joy in making things, even if they suck. Not everyone's going to like what you make, or agree with you, and that's why it's so important that you are one of your biggest fans. Nobody looks at my daily photo grid more than I do. I made all of that, and it feels so good. Obviously it feels good when other people tell me they like it too, but not hearing that doesn't discourage me.

And when you get started, send me a link and say hi so I can follow along. I can't tell you how important this has become to me (or maybe I just did), and I'd love to share that as widely as I can.

1. I hope/think that when the hype around AI dies down, what's going to remain is a massive improvement to searching for content online. The ability to effectively mathematically represent intent in language makes finding relevant results so much better. I just don't want the output to have any creativity or personality. Just find me the thing and show it to me. This won't be the end result, surely, but it's good to have dreams.

It is necessary to balance your survivor-like. It's not enough to just make a thing that sends waves of enemies, invent/borrow a load of weapons and stat boosts, then call it a day; these things need to be balanced.

I have been enjoying a YouTube channel for a while now, called Gohjoe. He just plays games, but almost exclusively the sorts of games that I like. He plays a lot of different games, and likes survivor-likes, too. Recently he's tried 2 that looked like a lot of fun - Yet Another Zombie Survivors and Jotunnslayer - so I bought them to have a go and they're both absolutely horribly balanced. I should not be able to win my very first round in one of these games.

In contrast, I've been playing Nimrods for months and I still haven't got past the third boss. I don't even know if there is one - I always die too soon. Same with Death Must Die; I've played so much of that game but I still haven't seen the end of it. I can never quite make it to the end. Other great ones are Megabonk (in spite of the absolutely terrible name), Halls of Torment, Brotato, Deep Rock Galactic Survivor, Gunsuit Guardians, Vampire Survivors (obviously) and Vampire Hunters.

The balancing is the difficult part, for sure, but it's also part that gives a game longevity. If I can finish a run on a roguelike on my first try, I am going to refund it.

There's so many of these games around right now and I'm very glad that Steam has such a good refund policy. I can't tell you how many I've tried and refunded. It's a lot.