I wish you good luck and many well copied models, ideas. 
oh look another one that doesn’t match
not only are the thumbprints different entirely, the ratio of the print to the pot is different, and the scale of the entire model is not misrepresented.
The first image, does not physically exist. So it looks like both of us are just doing our best to try and match it i guess?
I won’t be uploading any models, don’t worry, I won’t cut into your points mate XD
2 Likes
man this is a pretty ridiculous statement, these are not your ideas, it’s an ai model, you are also copying it yourself 
2 Likes
Really nice vid! There’s no need to be self-conscious about your voice - your voice is quite decent. The vid shows it’s non-public so only people with link can see it. Can you make it public for broader audience, please? This could also work as a great little tutorials for Fusion 360. If I knew FreeCAD as well I’d use that concept to do FreeCAD tutorials 
2 Likes
just fixed that then, noticed it was unlisted.
Regardless of my personal thoughts and emotions, it was genuinely enjoyable to take an image and try and recreate it. I’m no expert at it either
I don’t plan on doing anything other than that, I’m not profiting in any way and don’t intend to
2 Likes
A AI shows you the Idea that you have her. Not by itself.
man that’s just not true, the prompt is pretty basic… here you go:
you can replace the toothbrush for kitchen elements and it will spit out pretty much the same thing
3 Likes
You got a new subscriber. It’s not the same name as i have here but it’s a channel where i used to upload fun videos that i did with friends and gaming clips
About your voice, don’t even try and do it AI. It’s pretty good and the levels are fine. Maybe pass a filter as someone told already, but you don’t even need to.
I knew you were preparing something like this and it’s amazing, spared me the trouble of creating a dedicated channel like i was going to. Thank you once more.
I will be sharing it as much as possible because i have many friends and people in groups where i am too, that wanted a quick and concise video on how to do these kind of things. I hope you keep doing this and i can see your channel rising in views soon.
From here on forward is just Ai-nto the Moon 
Thank you for brightening my Sunday 
2 Likes
We must have different versions of Fusion, mine can’t do any of that 
. Seriously that was great, I subscribed. I’m just learning Fusion and those ribbed surfaces were something I really wanted to figure out. I thought your voice was natural and not forced. Well done.
1 Like
Yes you are right. It’s pretty simple, you just have to give the idea itself. So I have a lot of fodder for many, many new models. It’s hard to train your inventiveness. You have it or you don’t and you copy the ideas that others had. In any case, thanks again to everyone here for your attention.
Yeah this is a bit of an empty conversation, you confuse generalist and basic prompts with inventiveness, but anyone can do it very easily, and it’s fine if that’s what you want to do, but don’t claim that it’s your ideas, it really isn’t, and as mentioned before, it’s not intellectual property, therefore anyone that chooses to do it is not copying you. I have no interest in doing it myself personally, but anyone can really because it’s not inventive or original it’s just what a model spits out
edit/ and to illustrate my point:
super inventive i know
1 Like
What type of AI are you using? ChatGPT takes forever to spit out the right image (not for Makerworld).
By looking at the logo in the pictures, he used chatgpt. Maybe you are using the 4o mini model, and that’s the basic one. You can also use Bing/Copilot and it can create good pictures. But please!!! Don’t be one of those guys
Edit: you can also take a look at ComfyUI, that can be run locally on your machine, and then see what models you are interested in with Hugging Face and Civtai. You need a pretty beefy GPU to be able to take advantage of that and there are lots of models already trained for several use cases. Have fun
2 Likes
@jaayjeee Your title in the video you created has an error where it should read Thumbnail and instead it read Thimbnail 
Just to let you know. Shared it to the German and English groups where i am currently using and moderating and people noticed that error 
1 Like
I think it’s funny when we argue about the originality of stuff, when the subject of what we’re talking about is content generated by AI.
I mean, AI doesn’t have an original bone in it’s body. All it does is spit out amalgamations of what it’s trained on. These houseware collections are some of the most generic things I’ve seen in the 3d printing world. You could throw a stone and hit 20 designers doing this stuff. Now that more and more people are realizing you can generate concept AI images, this stuff even looks more the same than it use to. Like at least before there was some originally between designers. Now they all use AI to generate the same kind of images and produce the same kind of models.
8 Likes
New research published in Nature shows that the quality of the model’s output gradually degrades when AI trains on AI-generated data. As subsequent models produce output that is then used as training data for future models, the effect gets worse. [Source]
The more human designers abandon their jobs because of AI output, the more c_r_a_p AI will produce. The problem solves itself in the end…
4 Likes
That is absolutely true. As i said before, there will be a regression after the evolution. When a LLM (Large Language Model) is trained with a majority of already AI generated models, as already is in some cases, it will start outputting garbage. There are already some examples available. Go into any AI engine (chatgpt, copilot, gemini, etc.) and give them the same exact prompt and the output will be very similar. Even changing the prompt, the end result will be also almost the same with only variations on the background and objects.
Imagine an LLM being trained for cups only and imagine that same LLM crawling the internet and finding the models of some people here that have high download counts and so it gets them in the search engine easier. Now imagine the output that the same LLM will translate to you, in the real world. Since there is no variation in the crawled models at all, because they were already created using an AI, the output will be almost the same every time, without any innovation, as we can see already from most of the generated model pictures in some profiles. It’s just a cylinder, hollow and ribbed.
And those, my friends, are almost all the models we see everyday. Hollowed models with ribs that serve as a container for anything. A creator is fueled by creativity and the desire to make what seems impossible in our own eyes and when someone dumps cr@p everyday like a factory would do, that is not called creativity anymore but a desire to get money fast through other creators hard work and intelligence.
Wish you all a nice Sunday
1 Like
With only this you got to the point of this topic. Finally!
Good luck reporting in those cases. If it was that easy, most of the people that are mentioned here, and lots of those who are not mentioned but people know who they are, would have 99% of their models taken down. Is that easy huh?
Have you tried reporting one, just one, of the models here mentioned that do not correspond with the image? If not, then be my guest and try. Again, good luck.
Yesss… Go on…
That’s the whole point!
Or rather - the reports should be acted upon accordingly.
Sidenote on litho/hueforge
As for lithopanes and hueforges - most people don’t have much against them. I personally don’t mind them. Or rather - would mind them a lot less if I could easily filter them out Same thing with fidgets really - some people do love fidgets though, so I don’t want to see them gone, just it would be awesome if I could filter them out of my feed.
Deceptive images, such as ai gen covers that do not correspond to models and accounts using such practices, that’s the topic here. And obviously low-quality ai-gen models that use ai gen covers, so everything is all wrong… Those are bad for users and designers.
1 Like
Here are some suggestions while @MakerWorld is working on an overall strategy.
- Models or cover images suspected of using AI, either detected by MW’s own procedure or reported by users, will be put into a review queque pending review before making public.
- Models that are confirmed to be AI generated, either by MW’s review, or by users failing to provide proof that it’s made by human upon requested by MW review team, will have less exposure than those made by humans.
- Models generated by AI receive less reward points, such as 50% of the normal points, than models made by humans.
These rules will be temporary in nature, meant to provide a practical mitigation while a better solution is being worked on.
A better solution will be possible when the following happen:
- All major CAD and other software have implemented effective AI watermarking. This will make it easier to identify AI generate models or images and relies less on manual review.
- Applications or algorithms have been developed to separate high quality models from low quality ones. This will lessen the burden of manual review to determine the quality of models. Once it’s feasible to ensure quality using less manpower, how the models are made will no longer be important. Good models, regardless made by AI or by humans, will receive the same treatment. Bad models, either made by AI or by humans, will also be treated the same.