> We're on the cusp of a profound content shovelware crisis.
No, we've been inundated with low-quality content for decades now. This is nothing new.
Fortunately, ratings and reviews and popularity have always been an extremely effective antidote.
Even if 99% of stuff on a platform is total crap, nobody cares. It's a non-issue. Whether you're talking about music, books, TV shows, or whatever. The 1% rises to the top and you don't honestly need to pay attention to the rest.
If you choose to pay $20 for something that has 3 reviews that are all 1-star, then that's more your problem than the system's problem.
>No, we've been inundated with low-quality content for decades now. This is nothing new.
Trends matter too. If it gets to be an order of magnitude cheaper to shill your products on social media and forums, set up seo crap articles to phish users from search results, or churn out good old fashioned email scams, then expect an order of magnitude worse signal to noise ratio on the internet as a result. There could be a point reached where the internet is functionally broken, with the signal to noise ratio too low to make it useful for anything, save for navigating directly to known good hosts that themselves will become increasingly more lucrative targets for enshittification.
> No, we've been inundated with low-quality content for decades now. This is nothing new.
The problem with this argument is that scale matters.
When a small number of people were trading mp3 files on FTP servers it was not seen as a problem. When Napster came out, it was seen as a problem. It was correctly seen as a qualitative shift in the effect it would have on society.
I have to agree reviews work very well. But it distresses me that, knowing the absolute tsunami of garbage that awaits us without it, we are SO laissez faire when it comes to protecting that system. We allow companies to game reviews with kick backs, we lets spammers in posting fake reviews. This is currently our one wall of defense and cracks in it should terrify us.
We are also on the cusp of individual developers being able to produce works that used to take entire teams.
I'm working on simulating a small town using Generative AI agents, schedules, social interactions, realistic reactions to outside events, dialogue between characters, the whole shebang.
A year ago that wasn't an "after work side project".
I just did a full launch on https://www.generativestorytelling.ai/ - a side project that was only possible because of AI help. Between art assets and also coding in brand new areas that I hadn't used before, AIs are an obscene boost to what individuals can do.
The price and complexity of software development projects has been increasing for years now, AI is a huge reset on the amount of effort needed to make stuff.
I wouldn’t be able to trust running generative code that I don’t personally have context for, or hasn’t been reviewed by an experienced peer; even if it seems to work as intended. It’s the kind of irresponsible OceanGate mentality. As for blog posts, “better than nothing” doesn’t automatically make content “compelling”.
In this context: Generative ML models allow e.g. a single motivated writer with almost no budget to make a Visual Novel which they then could publish on Steam (before the policy change) for the world to see.
Write the script yourself, generate and curate 2D art assets, optionally generate and curate your OST / BGM, optionally generate and curate voice lines, put everything into Ren'Py. Done.
It's still very much not easy if you do not want to make shovelware, but it's possible now for a sole developer (or very small team) with no great artistic and musical talent.
I think game art for programmers or people with a game design idea but no visual arts chops is a very fitting use. As would be generating text for such a use, like dialog for an NPC.
> I think game art for programmers or people with a game design idea but no visual arts chops is a very fitting use. As would be generating text for such a use, like dialog for an NPC.
People aren't nearly creative enough with this.
You can use generative AI to give you a complete schedule for every NPC in the town. You can use generative AI to determine what the relationships between people in the town are.
I am working on using generative AI to simulate an entire small town.
I'm not an expert or especially well versed, but I feel like I've seen things from midjourney that could pass for semi-professional art, the kind that might be sold at a small time park exhibition on a sunny afternoon.
And with somebody super focused on figuring out the right prompts, maybe tweaking the images slightly, I think it could be workable in a game for sprites, textures, backgrounds.
We're on the cusp of low skill know nothings flooding the internet with low quality trash. AI produced content will not have value because you don't add anything meaningful to that process, you're entirely subordinated to the quirks of whatever models you're using. How do you compete in a sea of games that look exactly like yours because you're all using the same garbage to build the aesthetic value of your game?
>We're on the cusp of a profound content shovelware crisis.
Everywhere. Games, porn, text, articles, music, etc. For the generation that grew up with the internet already existing, this is their epoch moment, lives pre and post generative AIs.
I was thinking the other day that original artworks are going to be far more valued with a glut of AI generated.
Thinking slightly ahead, you can find your absolute favorite artist, and in seconds use their style you love so much to make the family portrait you would never be able to commission them to do. But going forward even more…
It’s just not the same as a print right? Well, thanks to AI being able to learn and determine where brush strokes would land, we take advancements from 3D printers and your desktop painting rig picks up a brush and paints it just as the artist would have.
Then going forward even more.. the artist himself needs some cash and knocks out 100 of these customs while they sleep, signs them, and now they are originals, sort of.
So… verifiable originals are going to be the hot thing. A painting with a video of the artist painting it… but not an AI generated video of course!
Maybe the artist will have to print it on location while you watch.
To me that implies that real animals are scarce; the price is not related to any kind of "artistic realness."
If they said that hand-crafted robotic snakes were more expensive than run-of-the-mill live-bred snakes, that would support your point about artistic realness.
> Well, thanks to AI being able to learn and determine where brush strokes would land, we take advancements from 3D printers and your desktop painting rig picks up a brush and paints it just as the artist would have.
Good luck convincing enough people in the art community to give you data on their process so you can do this. There's definitely enough out there in things like PSD files but the AI community has been so rude and antagonistic to art communities that most of them have a kneejerk hate reaction on any mention of the technology, and rightfully so. AI users and companies have been gleefully abusing artists from day 1.
Why would you assume it would need to be given? If you examine a painting with enough care, you could work out a few brush strokes. Now hand that to an AI that can process everything and come up with a style and how to paint it.
If you are making a thing, there is evidence of how it was made. Patten match enough of it and you’re done.
This has been an issue on Steam for ages, people call them "asset flip" games, because someone can buy a few $10 asset packs and piece together a game out of them.
AI generated content is not meaningfully better or worse than these low effort games, though taking the time to generate passable content with AI is probably a lot more effort than just using $50 worth of assets that are already packaged up for unity.
Asset store is still a more effective tool to create shovelware than AI. It's been this way for years. Recommendation systems (digital or otherwise) are already coping with it.
Automation will be a force multiplier for laziness and predation more so than for creativity.