CGTrader, a long-established online marketplace for 3D assets … is being flooded by AI-generated assets, representing one in six models — but they only account for only $1 out of every $90 in revenue.
The article is about a single marketplace for 3d models. Those numbers are stark and probably reflect people purposely not buying ai, as the article says. But this is a single website in a small niche of the economy. That’s just not at all “The Economy Has Spoken”.
I remember someone who make ads was noticing their engagement dropped to nil after “saving money” with AI ads… They did an experiment. They handmade a cheap ad with cardboard and humans.
Engagement went way back up. People crave real art in a world of slop.
I presume not many if so. But the article doesn’t include any examples beyond this one. The headline has vastly overblown what the article actually is. I know that’s pretty common now but it just bugs me.
Souvenir shops, maybe? I’ve been seeing lots of ugly AI slop printed on random objects in those corners and older tourists don’t seem to mind. Although that might change as soon as feedback from family and friends arrives.
Even beyond this little example, if people wanted an AI gen 3d asset they will AI gen themselves, no one gonna buy AI gen 3d asset because it has basically 0 value. The same thing probably happened to design company that shifted to AI gen. No one need to pay the AI design company to get an AI gen logo and pamphlet, they can do it themselves.
I am not sure the majority of genAI 3d models are even usable as is. 3d prints need a conscious approach that optimizes them for being printed bottom2top, with popular precision/nosle sizes options in mind, with more interesting build using that limitation to their advantage - see e.g. infill printing on chain-connected toys you probably saw sold in bulk on marketplaces - from free designs on the internet, not to mention springs and other material/technology-dependant tricks. 3d models for animation and wiring need to have clean topology to render without surprise glitches and wasting cpu time. And I’m not even talking about super serious CAD stuff that needs consistency and precision in the smallest details.
I had little time with 3d, but from what I had with vector images and auto tracing, broken or/and overcomplicated geometry can lead one to a point it’s just easier to redraw the thing than trying to fix it.
The article is about a single marketplace for 3d models. Those numbers are stark and probably reflect people purposely not buying ai, as the article says. But this is a single website in a small niche of the economy. That’s just not at all “The Economy Has Spoken”.
Headlines these days.😮💨
Is there a marketplace where lots of AI generated assets are selling?
I remember someone who make ads was noticing their engagement dropped to nil after “saving money” with AI ads… They did an experiment. They handmade a cheap ad with cardboard and humans.
Engagement went way back up. People crave real art in a world of slop.
Yes, and even moreso than before GenAI.
One doesn’t know what one takes for granted until one loses it, after all
I presume not many if so. But the article doesn’t include any examples beyond this one. The headline has vastly overblown what the article actually is. I know that’s pretty common now but it just bugs me.
Souvenir shops, maybe? I’ve been seeing lots of ugly AI slop printed on random objects in those corners and older tourists don’t seem to mind. Although that might change as soon as feedback from family and friends arrives.
Advertising software gaming just not literally the whole thing obviously
Even beyond this little example, if people wanted an AI gen 3d asset they will AI gen themselves, no one gonna buy AI gen 3d asset because it has basically 0 value. The same thing probably happened to design company that shifted to AI gen. No one need to pay the AI design company to get an AI gen logo and pamphlet, they can do it themselves.
I am not sure the majority of genAI 3d models are even usable as is. 3d prints need a conscious approach that optimizes them for being printed bottom2top, with popular precision/nosle sizes options in mind, with more interesting build using that limitation to their advantage - see e.g. infill printing on chain-connected toys you probably saw sold in bulk on marketplaces - from free designs on the internet, not to mention springs and other material/technology-dependant tricks. 3d models for animation and wiring need to have clean topology to render without surprise glitches and wasting cpu time. And I’m not even talking about super serious CAD stuff that needs consistency and precision in the smallest details.
I had little time with 3d, but from what I had with vector images and auto tracing, broken or/and overcomplicated geometry can lead one to a point it’s just easier to redraw the thing than trying to fix it.