See the subreddit for even more articles and discussion: https://www.reddit.c...MediaSynthesis/
It's time to discuss something I scarcely believed could exist until a couple years ago. In fact, part of me still refuses to believe this will happen. But if I were to give in to that delusion, I'd be denying reality.
What is that reality?
Artificial intelligence will create for us.
There's a lot I can talk about here, and I will touch on the more extreme part right now.
Yes. AI will be creative. Creativity is experience + abstract thought. Thus, the last thing humans excel at will no longer be our sole domain.
But I wanna talk about what happens when we fuse our creativities.
I wanna play a game. It's a game I've played before, but with new features. Let's take a classic— 1998's Sonic Adventure. One of the first video games I've ever played, and still one of my favorites. It's also the last Sonic the Hedgehog game I played, back in late 2013. Ever since, I've abandoned the franchise, but that is a whole other story.
My point is that I love the game. I also love its sequel. But there's no denying that it's a limited game. It's aged absolutely horribly, and the gameplay is questionable.
What if I wanted to remake the game, adding new features and functions? Let's say I wanted to take out almost every bottomless pit, instead creating what basically amounts to an open world game? I'd need hundreds thousands of dollars and dozens of artists and programmers to work for me. And even then, I'd only be capable of doing what I wanted if I kept the graphics at their same level, because the amount of content I'd need to add would be too much for any basic hard drive to handle.
But hey, there's a charm to the graphics, so I'd like to switch back and forth.
"Are you mad?"
Okay, so let's push it back to 2036 then. Now my PC can hold many exabytes of data. But that doesn't change anything. I still need to create the expanded game.
Well okay. It's 2036, right? I'll create it. How? With algorithms!
I'll tell the algorithm what to do; give it the game as a base, and describe in-depth what I want. Then let the algorithm create the pieces I want and put it all together. Now I can play that vastly expanded game. Sonic Adventure + Sonic Adventure 2, with all the crap I want added to it. First time in 22 years that I play a Sonic game.
Why stop at existing games? Why not completely new ones? Just describe the game. Describe the features. Describe the environments.
"But I can't write very well."
Then you don't have to. An algorithm can guide you through what you want and generate a list by itself. It still needs your input. It needs the most basic inputs, of course— do you want a first-person shooter, or a platformer, or a real-time strategy game? An MMORPG? Puzzle??
It'll even create the music, sound effects, and voice acting for the game. All sound is just patterns of sound waves. Generating any sound is possible. And of course, you can tell the algorithm exactly what you want and fine-tune it all yourself if you really need to.
It might take some time, but all you need is a master algorithm. You can create the equivalent of today's AAA games on your computer.
This works extremely well with Full Immersion Virtual Reality, when you wish to create whole worlds with your mind. But there's just a problem or two— what happens to all those put out of work by these algorithms? Hmm! That's what we've been debating for years.
"But this sounds impossible."
It's not. I can bring up an endless amount of proof that this AI infringement upon creative works is happening today. A great big example? Deep Dream. The latest iteration can take any picture and spin it into a painting of any style. I want a picture of Earth from Space in the style of the Impressionists. It can do that.
Or I could pay $500 for a person skilled in mimicking impressionism to do the same.
Which do you choose? I'd choose the first one in every situation except purposeful antemillennialism. Why? In the case of purposeful antemillennialism, a human doing such a thing is the point. But in other cases, I just want the picture. I don't care about the artistic labor put into it— I just want the art, I don't givva damn who or what creates it. We're still gonna pay $15 million for authentic pieces.
But if you think we can make a living off that, you're off your rocker. Why put in the time and effort if AI could do it faster and more cheaply? Because it's authentic and human, so you can make some money. But not much.
Video gaming will benefit from this. Imagine many millions of people sharing their algorithm-created games, often taking others' games and using algorithms to mod them. It's the culture we have today, but massively decentralized. But would the market still be so large? Probably not. You'd have a large number of free games and for those who get too greedy, they'd just get their games copied.
Then comes the music. Most people aren't musicians.
I love love love heavy rock, but I'm pisspoor with a guitar. Also, as a black guy, it would take a ridiculous amount of effort to grow my hair out and get it (and keep it) straightened. Also, I'm ugly— keep me away from a camera. And then there's all the realities of being in a musical group. Let's try using an algorithm. I can even get the Super 8mm video. 4 strapping young lads headbanging to ultraheavy '70s rock, being recorded with a vintage camera? Algorithms will provide.
Edited by Yuli Ban, 13 June 2019 - 10:46 PM.