HOME | DD

Description
It's been a while since I had posted landscapes. I love fantasy environments, but what I love the most is space. Ironically, I haven't made a lot of space-related pictures yet x). I've still got a fair amount, and I'll post them all here!

On this picture you can see Mars. Three of them, actually. Yes, NovelAI is stupid x). AI imagen often gives abstract results, because neural networks simply don't have consciousness. They don't have neurons either, by the way. Do you want to know their true nature?

A neural network is literally just a mathematical function. Like f(x)=3x+3. Except it has tens of thousands of variables instead of just a few. That mathematical function is trained on a database ; it means that each parameter is tweaked to get a specific value that fits the database.
That's the gist of it ^^.

All that to say that neural networks aren't able to tell what they are drawing, nor can they even think about it. That's why we get random pictures, like this one. It's also why it's able to make blatant errors like too much limbs on a character or misplaced blur. It just can't realize its mistakes!

In this case though, it's a happy accident.

Description
It's been a while since I had posted landscapes. I love fantasy environments, but what I love the most is space. Ironically, I haven't made a lot of space-related pictures yet x). I've still got a fair amount, and I'll post them all here!

On this picture you can see Mars. Three of them, actually. Yes, NovelAI is stupid x). AI imagen often gives abstract results, because neural networks simply don't have consciousness. They don't have neurons either, by the way. Do you want to know their true nature?

A neural network is literally just a mathematical function. Like f(x)=3x+3. Except it has tens of thousands of variables instead of just a few. That mathematical function is trained on a database ; it means that each parameter is tweaked to get a specific value that fits the database.
That's the gist of it ^^.

All that to say that neural networks aren't able to tell what they are drawing, nor can they even think about it. That's why we get random pictures, like this one. It's also why it's able to make blatant errors like too much limbs on a character or misplaced blur. It just can't realize its mistakes!

In this case though, it's a happy accident.