As far as you know, some content on this site was produced with the help of AI tools. Namely, most of the janky stuff. Here is a non-comprehensive list of content and the AI Tools responsible:
- Images: Most “pictures” were made with various Stable Diffusion 1.5 checkpoints in Automatic1111 with HiRes Fix, loopback scaler, and ControlNet. Those raw images are then made perfect in Adobe Photoshop, which also uses AI Tools in its Neural Filters. However, I will never admit to using Adobe Firefly generative fill for this site, you can’t make me. Oh, and I suppose I should mention a light use of face-swapping to make all Wonder Woman images look vaguely like Lynda Carter, but that should go without saying.
- Music: As of today, Oct. 10, 2023, all of the music used in OJ WOLFSMASHER DOT COM videos are either licensed through Canva, made with Suno.ai’s Chirp service, or made via something called Splash Pro. The ones that sound janky are all from Chirp, sorry. That’s why we’re moving away from it.
- Text-to-speech: The “storytime” videos use TTS from Bark, using randomly generated or built-in voices. The videos I make from OJ WOLFSMASHER DOT COM posts use janky TTS from NovelAI. Again, randomly generated or built-in voices. We have recently started using Coqui TTS in oobabooga. Ugh, those names are terrible.
- Text: A PORTION of the text in OJ WOLFSMASHER DOT COM posts is generated by AI tools, usually NovelAI’s wonderful and non-lobotomized “Kayra” model, or now, Mixtral. And even that small portion is greatly edited by a human. I’d say about 50% of the text is human-generated.
- Things we DON’T use: Products from those closed AI jerks at OpenAI, Anthropic, PI, and others. The only censored AI we use is Suno.ai’s Chirp service to make music, and that just means I can’t make the AI sing about things like CHUDs and losers.
Why do I use AI content? Two very, very good reasons:
This is a fake news site that makes fun of other news sites/content farms/Youtube channels/etc. All those things are rife with AI content, so it only makes sense that we would use it, too. It lends authenticity to the fake website.
But let’s say I went full non-synthetic content, just for funsies. The text would only take about twice as much work as it does now, to make up that remaining 25% of text. Not ideal, but doable.
The site’s images would have to be provided by some large corporation’s service, probably – giving me way less flexibility, and costing me way more money. Honestly, there’s no way I would, or could, do it. I could also pay an artist to make bespoke images for every post. That would cost me a LOT of time, flexibility, and money. And I’m not going to ask someone to do this for exposure in lieu of money.
As for the music, I actually know people who would be awesome at making music for this site. But I’m not going to ask them to do it for free, that would be stupid. Paying them would not be possible at this time.
I could pay voice actors instead of using TTS, I suppose. That would be great! Again, I don’t have money, and I don’t want to NOT pay them.
The only paid AI services I use for the site are: NovelAI (which I would subscribe to anyway) and Chirp ($10/month for 100 30-second songs). Canva is also starting to get into AI content, but I’m not really using Canva for that. The rest I run locally on my hardware, only paying The Man for the electricity it takes to generate things. So don’t @ me and say I’m just paying big tech companies instead of real people. STANDING OFFER: I will pay a good musical artist $10/month to make me fifty 1-minute-long made-to-order songs each month, in any genre I choose. Sound good? No? Wait, where are you going? Come back!
There are also deeper questions about generative AI, who controls it, the data it’s trained on, copyright, what art even is, etc. If you have an economics-minded or politics-minded binary stance against all AI-generated art, I regret to inform you that all art belongs to everyone, and that copyright is morally indefensible. Unless you think art should be gatekept from poor people because they don’t have enough money to purchase the right to experience it. Or maybe you also think that the process of making art should be gatekept behind artistic talent, and that no tools should exist to allow non-established-artists to make better art than they otherwise could. Is that what you think? Really?