Free 3D
3D Print Models
3D Scans
Animation & MoCap
Textures
Materials
Skin Reference
2D Game Assets
Sound Effects
Brushes & Tools
Reference Photos
Stock Images
HDR Images
3ds Max
Maya
Blender
Unreal Engine
Unity
ZBrush
Cinema 4D
LightWave
Marmoset
SketchUp
FBX
DAE
DXF
GLBThe slow creep in infiltration has started!
1,230Thread Activity
Luxe MuseSat, May 16
Luxe MuseSat, May 16
laurendevSat, May 16
diogenese19348Sat, May 16
Luxe MuseSat, May 16
Luxe MuseSat, May 16
diogenese19348Sat, May 16
laurendevSat, May 16
Luxe MuseSat, May 16
Luxe MuseSat, May 16
On both renderosity and here, we're seeing these little notifications on the product pages.
The slow infiltration of AI products has arrived.
The slow infiltration of AI products has arrived.
! REPORT
RenderHub has a "Hide AI" button with the other filter options. Just click that, and you won't see AI products in marketplace categories or searches.
I like that it remembers my preference. Just set it and forget it.
I like that it remembers my preference. Just set it and forget it.
REPLY
! REPORT
Tenserknot
Karma: 9,368
Mon, May 11Dont want to forget it, and I'm not an ostrich.
I like being aware of what's coming down the pipe.
I like being aware of what's coming down the pipe.
stevenjoseph8844
Karma: 2,577
Thu, May 14Yep! Mine is permanently clicked to hide AI products. I don't want to see that junk.
While I'm not great fan of AI image generation ( and have made my feelings clear previously on that ) I did buy the new 'Image to genesis 8/9 Pose generator' and I have to say I'm impressed. This is the sort of product that I think shows AI can actually help a DAZ user, rather than driving people away from it. ( I'm not affiliated with the seller, just a happy customer ).
I see it has an AI tag, which is fair since it uses AI to somehow create a pose file from a photo, but it would be sad if people missed it owing to having all AI products hidden.
I see it has an AI tag, which is fair since it uses AI to somehow create a pose file from a photo, but it would be sad if people missed it owing to having all AI products hidden.
REPLY
! REPORT
dirtrider00
Karma: 21,683
Mon, May 11I saw that product and stuck it into the wishlist. Kinda waiting to see some reviews on it but that would go really good for some of the photo reference photos I've seen on here, especially for the type of poses that aren't represented. (like I saw a photo pack for baseball and there just aren't any baseball pose packs available)
Wolf007
Karma: 163
Tue, May 12When i saw this yesterday i didn't pay enough attention and thought that it was just a tool that simple used the 'AI' in analizying the pictures one feeds. That would had been ok by me, a tool that may save you time.
Now, i was pointed out at one sentence in the description that changes everything: "This is an AI model trained on over 25,000 Daz Studio poses."
My understanding is that that case goes against all posible licenses in both in Daz store and even here because is a derivative work in direct competition with the originals.
Now, i was pointed out at one sentence in the description that changes everything: "This is an AI model trained on over 25,000 Daz Studio poses."
My understanding is that that case goes against all posible licenses in both in Daz store and even here because is a derivative work in direct competition with the originals.
diogenese19348
Karma: 420
Tue, May 12I think you may be reading things into that which aren't there. I'm pretty sure what didn't happen is he loaded in half the Zeddicuss catalog.
I bought the product, it does everything he says it does, and however pose making was done before, it's going to change. In fact, if I were in the motion capture business at any level I'd be taking notice. It's really quick, the system requirements are amazingly low, and it works from pretty much any resolution picture, I've had ones as low as 74 KB it was able to translate reasonably well. And it takes (I haven't timed it exactly) less than 10 seconds a pose generally.
My impression is when the people at DAZ3D find this they'll do a Man Friday on it, buy it out, and add it to the paid version of DAZ. My guess anyway.
This is the future of posing. People can kick and scream about it all they want, but getting a pose from a picture on to the rigging of a 3D model is now going to take much less time.
I bought the product, it does everything he says it does, and however pose making was done before, it's going to change. In fact, if I were in the motion capture business at any level I'd be taking notice. It's really quick, the system requirements are amazingly low, and it works from pretty much any resolution picture, I've had ones as low as 74 KB it was able to translate reasonably well. And it takes (I haven't timed it exactly) less than 10 seconds a pose generally.
My impression is when the people at DAZ3D find this they'll do a Man Friday on it, buy it out, and add it to the paid version of DAZ. My guess anyway.
This is the future of posing. People can kick and scream about it all they want, but getting a pose from a picture on to the rigging of a 3D model is now going to take much less time.
Wolf007
Karma: 163
Fri, May 15"I think you may be reading things into that which aren't there."
I'm reading what the description said: 'This is an AI model trained on over 25,000 Daz Studio poses.'
"I'm pretty sure what didn't happen is he loaded in half the Zeddicuss catalog."
You are right, even half Zeddicus catalog will probably not be enought, likely required more than half.
If this tool had had the 'Ai bit' be only the image processing part (and the rest just be regular program logic, or procedural) I would had been ok with it. I would had even consider it to not be tagged as 'Ai'. But the dev took shortcuts, instead of having a proper image procesed and translate it to joint controls; decided to just adapt a bridge and feed another AI portion (a generative one) to be trained with the work of others.
I don't care it gets the Job done (did not disputed it it). I care that is most likely in license violation of most of the products used to train this AI.
I'm totally ok with the idea of more universal tools. Not fan of the execution for this one in particular.
I'm reading what the description said: 'This is an AI model trained on over 25,000 Daz Studio poses.'
"I'm pretty sure what didn't happen is he loaded in half the Zeddicuss catalog."
You are right, even half Zeddicus catalog will probably not be enought, likely required more than half.
If this tool had had the 'Ai bit' be only the image processing part (and the rest just be regular program logic, or procedural) I would had been ok with it. I would had even consider it to not be tagged as 'Ai'. But the dev took shortcuts, instead of having a proper image procesed and translate it to joint controls; decided to just adapt a bridge and feed another AI portion (a generative one) to be trained with the work of others.
I don't care it gets the Job done (did not disputed it it). I care that is most likely in license violation of most of the products used to train this AI.
I'm totally ok with the idea of more universal tools. Not fan of the execution for this one in particular.
diogenese19348
Karma: 420
Fri, May 15What I meant was I don't think he took 25,000 vendor poses to do the training. There are a limited number of joints to work with,. and a limited number of rotations on each joint. If I was going to do the project I would change that data incrementally to generate the positions, then match that up with the output from the AI side. Trying to gather 25,000 pictures of PA poses to do that from would be a goofy way to go about it.
Which is what I meant in that first comment. "I think you may be reading things into that which aren't there"
I don't think he was misusing licenses. He's the only one that can answer that though.
Which is what I meant in that first comment. "I think you may be reading things into that which aren't there"
I don't think he was misusing licenses. He's the only one that can answer that though.
Luxe Muse
Karma: 7,226
Fri, May 15Wait until we find out that not only were creator's poses used, but also, not even paid for. Hello to the black market, where all our creations end up being pirated. The source of your new favorite little AI assistant tool.
diogenese19348
Karma: 420
Fri, May 15Folks, the programmer in me says what he did was divide those rotations into increments, 360 degree rotations, say 60 degree increments to start with, 6 poses per joint, whatever, make the files procedurally, create the pose with generic Genesis 8, compare to the output on the picture side. when he goes back, he halves the increment and creates another set, which is why the matching starts getting better.
Doesn't that make more sense to you then scarfing a bunch of vendor pictures then still having to come up with the matching duf file of rotations? Really??
And that is still training on DAZ poses, because those ARE DAZ poses. Just not vendor created ones.
Again, laurnedev is going to have to say, but that sure makes one heck of a lot more sense to me.
The problem is any time anybody here's the words "AI Training" they think IP theft. Thanks Sam Altman.
Doesn't that make more sense to you then scarfing a bunch of vendor pictures then still having to come up with the matching duf file of rotations? Really??
And that is still training on DAZ poses, because those ARE DAZ poses. Just not vendor created ones.
Again, laurnedev is going to have to say, but that sure makes one heck of a lot more sense to me.
The problem is any time anybody here's the words "AI Training" they think IP theft. Thanks Sam Altman.
Luxe Muse
Karma: 7,226
Fri, May 151. The seller will admit it... uh-huh...
2. Moreover, he "rotated every joint" and, I quote, "25,000 Daz Studio poses." Ahem, maybe he meant times...? I'm sure he made all of them, every single one, himself. How many artists here, over the years, have had 25,000 poses in the store, I wonder?
3. That is not how training an LLM works. Without poses that depict actual human movement, rather than simple rotations, the LLM will not learn how a human moves, stands, or anything like that. A good example would be how image-to-video or text-to-video LLMs are trained.
4. Hello to the programmer in you. I do not blame the AI. I blame the humans.
2. Moreover, he "rotated every joint" and, I quote, "25,000 Daz Studio poses." Ahem, maybe he meant times...? I'm sure he made all of them, every single one, himself. How many artists here, over the years, have had 25,000 poses in the store, I wonder?
3. That is not how training an LLM works. Without poses that depict actual human movement, rather than simple rotations, the LLM will not learn how a human moves, stands, or anything like that. A good example would be how image-to-video or text-to-video LLMs are trained.
4. Hello to the programmer in you. I do not blame the AI. I blame the humans.
Luxe Muse
Karma: 7,226
Sat, May 16Is important to note, I'm not saying what he did was as easy, creating a dataset takes some effort... but compared to the time it took those creators that made the 25000 poses used for his dataset, it was a walk in the part...
diogenese19348
Karma: 420
Sat, May 16I can generate 25,000 .duf pose files with a script. You can to. The question is how the 25,000 pose files were acquired. That's what would make them legit vs stolen.
What I am saying is that if you were trying to match up data from an AI analysis to get rotations for a duf file it would be far faster to generate organized target files with a script then to try and match them up against vendor files. Really.
The key here is the program needs to translate what comes out of the AI pictures side into a set of rotations, ie:
{
"file_version" : "0.6.1.0",
"asset_info" : {
"id" : "/%21Poses%20%28DAZ%29/Genesis%208/Poses/%21AI%20Poses/On%20Floor/Floor%2001.duf",
"type" : "preset_pose",
"contributor" : {
"author" : "diogenese19348",
"email" : "",
"website" : ""
},
"revision" : "1.0",
"modified" : "2026-05-15T08:55:20Z"
},
"scene" : {
"animations" : [
{
"url" : "name://@selection/lSmallToe4_2:?translation/x/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?translation/y/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?translation/z/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?rotation/x/value",
"keys" : [ [ 0, -5.909195 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?rotation/y/value",
"keys" : [ [ 0, 0 ] ]
},
Trying to set that up against 25,000 random vendor files would be in a word, stupid.
What I am saying is that if you were trying to match up data from an AI analysis to get rotations for a duf file it would be far faster to generate organized target files with a script then to try and match them up against vendor files. Really.
The key here is the program needs to translate what comes out of the AI pictures side into a set of rotations, ie:
{
"file_version" : "0.6.1.0",
"asset_info" : {
"id" : "/%21Poses%20%28DAZ%29/Genesis%208/Poses/%21AI%20Poses/On%20Floor/Floor%2001.duf",
"type" : "preset_pose",
"contributor" : {
"author" : "diogenese19348",
"email" : "",
"website" : ""
},
"revision" : "1.0",
"modified" : "2026-05-15T08:55:20Z"
},
"scene" : {
"animations" : [
{
"url" : "name://@selection/lSmallToe4_2:?translation/x/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?translation/y/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?translation/z/value",
"keys" : [ [ 0, 0 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?rotation/x/value",
"keys" : [ [ 0, -5.909195 ] ]
},
{
"url" : "name://@selection/lSmallToe4_2:?rotation/y/value",
"keys" : [ [ 0, 0 ] ]
},
Trying to set that up against 25,000 random vendor files would be in a word, stupid.
laurendev
Karma: 445
Sat, May 16It is probably best to just clear the air so people know exactly how it works under the hood. The whole issue of AI being trained on other people's data is a massive topic right now. Even giant companies like OpenAI are dealing with lawsuits over it every single day.
To be completely transparent, my original plan actually was to train the AI using existing pose packs. But I ended up scrapping that idea for a couple of reasons.
First, as you know, pose presets always come with a preview thumbnail (.png). To train the model, it needs both the visual reference and the .duf file. The problem is that the resolution on those commercial preview thumbnails is incredibly low. Using them would have actually hurt the training process rather than helping it, because the AI would not be able to clearly make out what it was looking at.
Second, and this ties exactly into what the forum is worried about, I just was not sure about the ethics of doing it that way.
Because of that, I took a completely different route. I wrote my own custom Daz Script to automate the entire process from scratch. For example, the script would pose a character by just raising its right arm, and then take screenshots from 7 different angles. Then it would move on to another pose, like raising a leg, and do the same thing.
It just kept running through permutations until I had exactly 25,000 unique pose files, complete with image files from 7 different angles for every single pose. Doing it this way is what allowed me to build such an exceptionally clean and reliable training dataset.
To be completely transparent, my original plan actually was to train the AI using existing pose packs. But I ended up scrapping that idea for a couple of reasons.
First, as you know, pose presets always come with a preview thumbnail (.png). To train the model, it needs both the visual reference and the .duf file. The problem is that the resolution on those commercial preview thumbnails is incredibly low. Using them would have actually hurt the training process rather than helping it, because the AI would not be able to clearly make out what it was looking at.
Second, and this ties exactly into what the forum is worried about, I just was not sure about the ethics of doing it that way.
Because of that, I took a completely different route. I wrote my own custom Daz Script to automate the entire process from scratch. For example, the script would pose a character by just raising its right arm, and then take screenshots from 7 different angles. Then it would move on to another pose, like raising a leg, and do the same thing.
It just kept running through permutations until I had exactly 25,000 unique pose files, complete with image files from 7 different angles for every single pose. Doing it this way is what allowed me to build such an exceptionally clean and reliable training dataset.
laurendev
Karma: 445
Sat, May 16On top of that, 25,000+ poses is a massive number. Scraping together over 25,000 pre-made pose packs is no easy feat. And even if I did manage to collect them all, how could I guarantee that every single one of them would be flawless? If I had used 25,000 ready-made poses, there would inevitably be duplicates, or poses that end up throwing the character completely out of frame when applied. There was just no way I could manually check 25,000 poses one by one. When you combine all of this (along with the issues I mentioned above), it would actually do a lot more harm than good to the AI.
That's exactly why I wrote my own script to take care of the manual adjustments. If anyone is skeptical about this, just shoot me a DM and I'll happily send you the .dsa script I wrote
That's exactly why I wrote my own script to take care of the manual adjustments. If anyone is skeptical about this, just shoot me a DM and I'll happily send you the .dsa script I wrote
laurendev
Karma: 445
Sat, May 16I could have just looked past the issues I mentioned above, or found some workaround, and still used 25,000 pre-made pose packs.
But the real dealbreaker for me was the absolute mess that kind of dataset would create. Picture this: the first 10 files might be walking poses, and the next 10 could be sleeping poses. The character would be standing up, and then instantly snap into a drastically different position.
Essentially, mashing together 25,000 pre-made poses would just result in a chaotic, inconsistent dataset, which would completely confuse the AI. It can't start learning standing poses and then abruptly get thrown into something totally unrelated. The model needed to learn sequentially, with smooth, logical transitions between poses, rather than making massive, jarring jumps.
Now, to address the accusations being thrown around: The line in the product description that says 'trained on over 25,000 Daz Studio poses' does not mean 'trained on 25,000 commercial pose packs.' I could have easily just omitted that sentence and kept the inner workings of the tool completely under wraps.
You can't just hurl accusations about a tool when you aren't sure how it actually functions under the hood. It's one thing to have a debate, but it's another to make baseless accusations, especially about a technical process you clearly don't fully understand.
You don't have to like the tool. You are completely free to hate it. But what you cannot do is go on a public forum and falsely accuse me of piracy, license violations, and training my model on stolen, unpaid assets.
And hey, if building a tool like this using 25,000 pre-made poses is really just a 'walk in the park' like you claim, then by all means, go ahead and build a better one. Nobody is stopping you
But the real dealbreaker for me was the absolute mess that kind of dataset would create. Picture this: the first 10 files might be walking poses, and the next 10 could be sleeping poses. The character would be standing up, and then instantly snap into a drastically different position.
Essentially, mashing together 25,000 pre-made poses would just result in a chaotic, inconsistent dataset, which would completely confuse the AI. It can't start learning standing poses and then abruptly get thrown into something totally unrelated. The model needed to learn sequentially, with smooth, logical transitions between poses, rather than making massive, jarring jumps.
Now, to address the accusations being thrown around: The line in the product description that says 'trained on over 25,000 Daz Studio poses' does not mean 'trained on 25,000 commercial pose packs.' I could have easily just omitted that sentence and kept the inner workings of the tool completely under wraps.
You can't just hurl accusations about a tool when you aren't sure how it actually functions under the hood. It's one thing to have a debate, but it's another to make baseless accusations, especially about a technical process you clearly don't fully understand.
You don't have to like the tool. You are completely free to hate it. But what you cannot do is go on a public forum and falsely accuse me of piracy, license violations, and training my model on stolen, unpaid assets.
And hey, if building a tool like this using 25,000 pre-made poses is really just a 'walk in the park' like you claim, then by all means, go ahead and build a better one. Nobody is stopping you
Wolf007
Karma: 163
Sat, May 16"'trained on over 25,000 Daz Studio poses' does not mean 'trained on 25,000 commercial pose packs.'"
It also does not mean 'trained on 25.000 procedural generated poses with 7 renders each'. Your wording left it open to interpretation,
We are talking of 'AI training' that has FACTUAL HISTORY of using others people's work without permision. So I'm not going to apologize for expressing my doubts; your wording on the 25k, your mistake.
And you don't have to DM me anything to prove your claims, i'm just a random and also inclined to belive you; but do have in mind that your product has already been known to Daz with some people having interpreted your 25k as potentially using commercial pre-made poses. It wasn't me, someone was promoting your product in their forums, thread deleted obviously but i doubt Daz will leave it at that.
It also does not mean 'trained on 25.000 procedural generated poses with 7 renders each'. Your wording left it open to interpretation,
We are talking of 'AI training' that has FACTUAL HISTORY of using others people's work without permision. So I'm not going to apologize for expressing my doubts; your wording on the 25k, your mistake.
And you don't have to DM me anything to prove your claims, i'm just a random and also inclined to belive you; but do have in mind that your product has already been known to Daz with some people having interpreted your 25k as potentially using commercial pre-made poses. It wasn't me, someone was promoting your product in their forums, thread deleted obviously but i doubt Daz will leave it at that.
Luxe Muse
Karma: 7,226
Sat, May 16@lauerndev:
1. Using common sense is not hurling accusations.
2. I have not read anywhere that anyone mentioned this being a "walk in the park". I even said "it takes effort," and now I will add that it takes a little more effort than what you took to read the comment.
3. I do not know your product, nor did I go to your product page. I simply read "25,000 Daz Studio poses" in this same forum and realized, using common sense, how much effort it would take to create 25,000 poses.
I do not mind if AI corrects my typos or fixes my ESL mistakes when I need it. However, I do mind when humans step over other people's hard work, and I will forever remain skeptical.
See, I follow an artist here who has been making a replica of a spaceship. It has been almost a year, if not more. I have lost count. I admire his patience, and I do not doubt his craft. I follow other artists here whom I look up to. They make clothing, scenery, characters, or simply renders. They sometimes share their challenges, ideas, and even teach others. It hurts to know that at any time someone may take the food off their table, not through fair competition, but by using their own creations and hard work.
Also, we can all use AI to ask how you made the tool. It is funny how that works. But just like no one said it was a walk in the park, I have no idea where you thought we wanted to make a better one.
@diogenese19348 "Trying to set that up against 25,000 random vendor files would be in a word, stupid."
It sure is stupid to simply use the 256x256 tip files and ask Qwen to generate the caption, or to use the AI Toolkit auto-caption to generate the dataset...
In fact, an even more stupid thing, to put it in a word, is that you can ask Claude to build the auto-dataset creator that will walk through your pose directory, find the tip files, and build the JSON dataset for you that describes the pose in the duf file...
All that is probably as stupid as crafting 25,000 poses for an AI tool faster than many artists who have spent years producing poses here.
1. Using common sense is not hurling accusations.
2. I have not read anywhere that anyone mentioned this being a "walk in the park". I even said "it takes effort," and now I will add that it takes a little more effort than what you took to read the comment.
3. I do not know your product, nor did I go to your product page. I simply read "25,000 Daz Studio poses" in this same forum and realized, using common sense, how much effort it would take to create 25,000 poses.
I do not mind if AI corrects my typos or fixes my ESL mistakes when I need it. However, I do mind when humans step over other people's hard work, and I will forever remain skeptical.
See, I follow an artist here who has been making a replica of a spaceship. It has been almost a year, if not more. I have lost count. I admire his patience, and I do not doubt his craft. I follow other artists here whom I look up to. They make clothing, scenery, characters, or simply renders. They sometimes share their challenges, ideas, and even teach others. It hurts to know that at any time someone may take the food off their table, not through fair competition, but by using their own creations and hard work.
Also, we can all use AI to ask how you made the tool. It is funny how that works. But just like no one said it was a walk in the park, I have no idea where you thought we wanted to make a better one.
@diogenese19348 "Trying to set that up against 25,000 random vendor files would be in a word, stupid."
It sure is stupid to simply use the 256x256 tip files and ask Qwen to generate the caption, or to use the AI Toolkit auto-caption to generate the dataset...
In fact, an even more stupid thing, to put it in a word, is that you can ask Claude to build the auto-dataset creator that will walk through your pose directory, find the tip files, and build the JSON dataset for you that describes the pose in the duf file...
All that is probably as stupid as crafting 25,000 poses for an AI tool faster than many artists who have spent years producing poses here.
diogenese19348
Karma: 420
Sat, May 16He wasn't crafting poses, he was measuring a series of joint rotations, but you just go on being you. Seriously. I'm talking programming issues and you are hung up on AI. He told you why he did it the way he did. I figured that was the most efficient way to do it. But if you want to ignore all that and continue the argument, be my guest. I'm done with it.
Luxe Muse
Karma: 7,226
Sat, May 16You were trying to explain the pretzel. As I said before, to train an LLM, you need to provide good data and I quote: "Without poses that depict actual human movement, rather than simple rotations, the LLM will not learn how a human moves, stands, or anything like that. "
That means no random auto-generated joint rotations. I am, however, focusing on the "how does it achieve the pose" part of the tool.
I do understand that in order for the tool to interact with DAZ, it also has to learn the SDK and so on. But the only thing I care about is the art. We are in Renderhub, not LinkedIn.
To draw a parallel, you do not train an LTX 2.3 LoRA on people smiling by feeding videos of ants carrying sticks.
That means no random auto-generated joint rotations. I am, however, focusing on the "how does it achieve the pose" part of the tool.
I do understand that in order for the tool to interact with DAZ, it also has to learn the SDK and so on. But the only thing I care about is the art. We are in Renderhub, not LinkedIn.
To draw a parallel, you do not train an LTX 2.3 LoRA on people smiling by feeding videos of ants carrying sticks.
laurendev
Karma: 445
Sat, May 16You just admitted that you haven't even looked at the product or the product page. That honestly tells me everything I need to know here. Accusing someone of piracy, theft, and 'taking food off artists' tables' over a tool you haven't even bothered to click on isn't 'common sense', it's just reckless.
As for your elaborate theory about me using Claude or Qwen to auto-caption 256x256 tip files... again, if you had actually read my very first post, you would know I explicitly stated that I scrapped the idea of using thumbnails entirely. You clearly have no idea just how useless a 256x256 resolution is for an AI trying to accurately learn a pose, yet you continue to speak as if you're a technical expert on the matter.
My custom script didn't scrape a directory of pirated .duf files or vendor thumbnails. It mathematically rotated joints on a blank character right in the viewport and took high-res screenshots from 7 different camera angles. There were no pre-made vendor files used. Period.
Playing armchair detective and fabricating technical scenarios about a product you just admitted you haven't even looked at doesn't help anyone.
Let's be clear: what you are doing right now isn't 'using common sense', it's outright slander. And you're doing it without even having the slightest clue about what you're actually talking about. If you ever decide to look at the facts, the offer to review my .dsa script still stands. Until then, there's really nothing left to discuss.
As for your elaborate theory about me using Claude or Qwen to auto-caption 256x256 tip files... again, if you had actually read my very first post, you would know I explicitly stated that I scrapped the idea of using thumbnails entirely. You clearly have no idea just how useless a 256x256 resolution is for an AI trying to accurately learn a pose, yet you continue to speak as if you're a technical expert on the matter.
My custom script didn't scrape a directory of pirated .duf files or vendor thumbnails. It mathematically rotated joints on a blank character right in the viewport and took high-res screenshots from 7 different camera angles. There were no pre-made vendor files used. Period.
Playing armchair detective and fabricating technical scenarios about a product you just admitted you haven't even looked at doesn't help anyone.
Let's be clear: what you are doing right now isn't 'using common sense', it's outright slander. And you're doing it without even having the slightest clue about what you're actually talking about. If you ever decide to look at the facts, the offer to review my .dsa script still stands. Until then, there's really nothing left to discuss.
laurendev
Karma: 445
Sat, May 16And since we're on the topic of how AI actually works, I can't let this slide either: You are fundamentally misunderstanding what this tool even does.
The program doesn't need to know what 'actual human movement' looks like inside Daz Studio. It doesn't need to be trained on perfect, anatomically correct human poses. Why? Because it isn't generating poses out of thin air. You aren't typing in a prompt saying, 'generate a walking pose.'
The AI already knows what the human pose looks like because you are feeding it a real reference photo. The photo provides the anatomy. The only thing the model actually needs to learn is the translation layer: how to physically manipulate and map that visual data onto Daz Studio's specific joint rotation dials. It just needs to understand Daz's mechanical logic.
So your 'ants carrying sticks' analogy makes zero sense here. A more accurate analogy would be teaching a driver how the steering wheel in a new car works. They already know how to drive; they just need to learn the specific controls. That is exactly why a dataset of mathematical joint rotations was the correct way to train it
The program doesn't need to know what 'actual human movement' looks like inside Daz Studio. It doesn't need to be trained on perfect, anatomically correct human poses. Why? Because it isn't generating poses out of thin air. You aren't typing in a prompt saying, 'generate a walking pose.'
The AI already knows what the human pose looks like because you are feeding it a real reference photo. The photo provides the anatomy. The only thing the model actually needs to learn is the translation layer: how to physically manipulate and map that visual data onto Daz Studio's specific joint rotation dials. It just needs to understand Daz's mechanical logic.
So your 'ants carrying sticks' analogy makes zero sense here. A more accurate analogy would be teaching a driver how the steering wheel in a new car works. They already know how to drive; they just need to learn the specific controls. That is exactly why a dataset of mathematical joint rotations was the correct way to train it
Luxe Muse
Karma: 7,226
Sat, May 16Thumbnails would be for creating captions, not as the source. I also do not need to buy or see your product if the key point I am arguing about is the factual "happy customer" comment around 25,000 poses that was shared verbatim here.
Things can get a lot easier if you are following this thread. You can simply ask your favorite AI. Drop this prompt:
If you had to train an LLM on how to make DAZ Studio characters pose in a certain way, what type of data would you feed it during training? Please think hard about this.
Things can get a lot easier if you are following this thread. You can simply ask your favorite AI. Drop this prompt:
If you had to train an LLM on how to make DAZ Studio characters pose in a certain way, what type of data would you feed it during training? Please think hard about this.
Luxe Muse
Karma: 7,226
Sat, May 16Also... hi, hi, hi... 7 different camera angles. on 25000 poses... boy would that take time to render... please, stop insulting everyone here.
And good, I'm guessing your script could do this, you are a genius
And good, I'm guessing your script could do this, you are a genius
laurendev
Karma: 445
Sat, May 16Let's be direct, because at this point you've earned it.
You called my work "black market," accused me of selling tools built on "stolen, unpaid assets," and said I'm "taking food off artists' tables." Those aren't opinions, those aren't "common sense," they are accusations of theft and piracy, made in public, against a real person with a real product. And you made every single one of them while admitting, in your own words, "I do not know your product, nor did I go to your product page."
Read that back to yourself. You publicly accused someone of running a piracy operation, and in the same thread you admitted you never even clicked on the product page. That isn't skepticism. That's slander, full stop.
Now let's clean up the technical mess you keep making.
"7 angles 25,000 poses, boy would that take time to render." Nothing was rendered. The script grabbed viewport snapshots through Daz's scripting API, around 100 ms each, ray-tracing not involved. It ran overnight. You don't get to lecture anyone about "how training data works" when you can't tell the difference between a render and a screenshot
"Without poses that depict actual human movement, the LLM will not learn how a human moves." This is not an LLM. It is not a generative model. It is an image-conditioned regression head: photo in, joint-rotation values out. The reference photo carries the anatomy. The model only learns the mapping from visual landmarks to Daz's rig controls. Your "ants carrying sticks" analogy describes a problem that does not exist in this architecture. If you'd looked at the product page even once, you'd know that
"You are a genius." Save the sarcasm. The .dsa script that built the entire training set is sitting on my drive, and I have offered it publicly, twice. You haven't asked for it. Neither has anyone else who's been the loudest in this thread. Anyone who actually wants to verify how the data was generated can DM me and I'll send it within the hour. The offer is open and it stays open
What you are doing isn't critique, isn't debate, isn't "using common sense." It is making serious accusations about a product you've publicly stated you haven't looked at, and then moving the goalposts every time the technical reality contradicts you. First it was stolen vendor files, then it was scraped thumbnails, now it's a "happy customer comment." Pick a story.
One last piece of advice, since you've been so free with yours. If you don't know how something works, don't speak about it as if you do. Confidently lecturing in a public forum about a subject you haven't taken the time to understand, then dressing it up as "common sense," is how reputations get damaged. Yours and other people's. People remember who showed up to learn and who showed up to throw stones
I'm done addressing you specifically. If anyone reading this wants the script, the dataset structure, or a real technical breakdown of how the tool works, my DMs are open. To everyone else here who's actually engaged in good faith, thank you, sincerely
You called my work "black market," accused me of selling tools built on "stolen, unpaid assets," and said I'm "taking food off artists' tables." Those aren't opinions, those aren't "common sense," they are accusations of theft and piracy, made in public, against a real person with a real product. And you made every single one of them while admitting, in your own words, "I do not know your product, nor did I go to your product page."
Read that back to yourself. You publicly accused someone of running a piracy operation, and in the same thread you admitted you never even clicked on the product page. That isn't skepticism. That's slander, full stop.
Now let's clean up the technical mess you keep making.
"7 angles 25,000 poses, boy would that take time to render." Nothing was rendered. The script grabbed viewport snapshots through Daz's scripting API, around 100 ms each, ray-tracing not involved. It ran overnight. You don't get to lecture anyone about "how training data works" when you can't tell the difference between a render and a screenshot
"Without poses that depict actual human movement, the LLM will not learn how a human moves." This is not an LLM. It is not a generative model. It is an image-conditioned regression head: photo in, joint-rotation values out. The reference photo carries the anatomy. The model only learns the mapping from visual landmarks to Daz's rig controls. Your "ants carrying sticks" analogy describes a problem that does not exist in this architecture. If you'd looked at the product page even once, you'd know that
"You are a genius." Save the sarcasm. The .dsa script that built the entire training set is sitting on my drive, and I have offered it publicly, twice. You haven't asked for it. Neither has anyone else who's been the loudest in this thread. Anyone who actually wants to verify how the data was generated can DM me and I'll send it within the hour. The offer is open and it stays open
What you are doing isn't critique, isn't debate, isn't "using common sense." It is making serious accusations about a product you've publicly stated you haven't looked at, and then moving the goalposts every time the technical reality contradicts you. First it was stolen vendor files, then it was scraped thumbnails, now it's a "happy customer comment." Pick a story.
One last piece of advice, since you've been so free with yours. If you don't know how something works, don't speak about it as if you do. Confidently lecturing in a public forum about a subject you haven't taken the time to understand, then dressing it up as "common sense," is how reputations get damaged. Yours and other people's. People remember who showed up to learn and who showed up to throw stones
I'm done addressing you specifically. If anyone reading this wants the script, the dataset structure, or a real technical breakdown of how the tool works, my DMs are open. To everyone else here who's actually engaged in good faith, thank you, sincerely
Luxe Muse
Karma: 7,226
Sat, May 16This will be the last time I address you directly as well, because the more you talk, the less I think about the AI tool and the more I think about the person and intentions behind it.
You wrote on your product page that you trained your model on 25,000 poses. Worse, now you said you produced seven (7, not even one) high-definition renders for each pose to train your LLM. 25000 * 7
Had you said 100 poses, people, me included, would have said, “Wow, that is some work.Even 200, “Wow, that is serious work.But 25,000?
It is unfortunate that others will learn to be a little more deceptive after your mistakes. But it also means that what little trust we had in people “using AIat a place like this is probably completely shattered.
We know it is impossible for someone to produce 25,000 poses that could properly teach an AI how to pose a character. Your automation excuse about making a script that creates accurate human poses does not hold up either.
If it is true (just the script that poses), that in itself would have been “THE PRODUCT,because not even great programmers like Zeb0 in the DAZ ecosystem have made such a tool.
1. Using common sense, I can deduce that you did not make 25,000 poses.
2. Using common sense, those 25,000 poses must have come from somewhere.
3. I have no evidence of your personal wealth. Buying that amount of poses does not come cheap.
We all know about piracy here. We have all fallen victim to it. In fact, I am sure we all check from time to time to see what the latest product leaked by bad actors is. It is not too far-fetched to deduce that a possible way of acquiring 25,000 poses is through those channels.
You wrote on your product page that you trained your model on 25,000 poses. Worse, now you said you produced seven (7, not even one) high-definition renders for each pose to train your LLM. 25000 * 7
Had you said 100 poses, people, me included, would have said, “Wow, that is some work.Even 200, “Wow, that is serious work.But 25,000?
It is unfortunate that others will learn to be a little more deceptive after your mistakes. But it also means that what little trust we had in people “using AIat a place like this is probably completely shattered.
We know it is impossible for someone to produce 25,000 poses that could properly teach an AI how to pose a character. Your automation excuse about making a script that creates accurate human poses does not hold up either.
If it is true (just the script that poses), that in itself would have been “THE PRODUCT,because not even great programmers like Zeb0 in the DAZ ecosystem have made such a tool.
1. Using common sense, I can deduce that you did not make 25,000 poses.
2. Using common sense, those 25,000 poses must have come from somewhere.
3. I have no evidence of your personal wealth. Buying that amount of poses does not come cheap.
We all know about piracy here. We have all fallen victim to it. In fact, I am sure we all check from time to time to see what the latest product leaked by bad actors is. It is not too far-fetched to deduce that a possible way of acquiring 25,000 poses is through those channels.
Wolf007
Karma: 163
Sat, May 16I think i was the one who first said 'render' about the 7 images. My bad, sorry if that lead someone the wrong way.
Luxe Muse
Karma: 7,226
Sat, May 16@Wolf007
He said: "My custom script didn't scrape a directory of pirated .duf files or vendor thumbnails. It mathematically rotated joints on a blank character right in the viewport and took high-res screenshots from 7 different camera angles. There were no pre-made vendor files used. Period."
He said: "My custom script didn't scrape a directory of pirated .duf files or vendor thumbnails. It mathematically rotated joints on a blank character right in the viewport and took high-res screenshots from 7 different camera angles. There were no pre-made vendor files used. Period."
Luxe Muse
Karma: 7,226
Sat, May 16We need an edit button 
Anyways, an LLM trained with mathematically rotated joints would learn to produce joint rotations. But it wouldn't understand a request like "Have my character in a sexy crouch pose, smiling and winking."
Anyways, an LLM trained with mathematically rotated joints would learn to produce joint rotations. But it wouldn't understand a request like "Have my character in a sexy crouch pose, smiling and winking."
Luxe Muse
Karma: 7,226
Sat, May 16Guess, I was force to give a "free view" to the product page. It even states "Trained on 25,000+ pose files. ~80-85% accuracy."
Luxe Muse
Karma: 7,226
Sat, May 16"- PG4 Capybara (April - Current): Uses a two-tier correction system. A primary AI estimates the general pose, and then 107 micro-models step in to individually correct the error for each specific bone. I even added "expert" models for the 12 channels it struggles with most. Trained on 25,000+ pose files. ~80-85% accuracy."
laurendev
Karma: 445
Sat, May 16You keep arguing against a tool that does not exist.
For the third time now. The script never produced 25,000 anatomically correct human poses. It produced 25,000 mathematical permutations of joint rotations on a blank Genesis 8 figure. Right arm raised 30 degrees, then 60, then 90. Left knee bent 45 degrees, combined with shoulder twisted 20 degrees. And so on, programmatically, through the rotation space of the rig. Most of those poses look weird. Many of them look impossible. That is the entire point.
You just wrote, word for word: "We know it is impossible for someone to produce 25,000 poses that could properly teach an AI how to pose a character." Correct. It would be impossible, and nobody tried. What I built is not a model that learned "what human poses look like." It is a model that learned how the Daz rig responds to inputs. A mapping table between a reference image and the rotation dials. The anatomy comes from the photo you feed it at inference time, not from the training data. The training set does not need to contain a single realistic human pose for the model to work, and it doesn't.
I told you this in my very first reply in this thread. I told you again in my last reply, in plain English: "It doesn't need to be trained on perfect, anatomically correct human poses. Why? Because it isn't generating poses out of thin air." You ignored it both times and went straight back to "impossible to produce 25,000 real poses," which is an argument against a tool I never described and never built. If you keep swinging at that strawman, that is on you, not on me.
And while we are on the subject of things you have ignored. I have offered the .dsa script that built the entire dataset at least three times now, openly, in this thread. Anyone who wants to see exactly how the data was generated can DM me and I will send it within the hour. Not you specifically, anyone. You haven't asked. None of the people making the loudest accusations in this thread have asked. You would rather speculate publicly than spend 60 seconds verifying the thing you are speculating about. That tells the audience here everything they need to know about whether this is a good-faith conversation.
While we are here, the three new accusations you packed into that post.
"High-definition renders." Each sample in the dataset is one .duf file plus seven viewport snapshots, 969x729 PNG, around 110 KB each. Those are the file properties of viewport screenshots, not ray-traced renders. A real HD render at that resolution, with lighting and materials, would be many times larger and take hundreds of times longer to produce. I can post the file properties of any sample on request. You won't ask, because verifying the claim would end the line of attack.
"Train your LLM." This is not an LLM. I told you that explicitly in my last reply and explained what it actually is, an image-conditioned regression head that maps a reference photo to joint-rotation values. You repeated the wrong term anyway. At this point that is a rhetorical choice, not a misunderstanding, and I think you know it. Calling it an LLM lets you keep arguing as if the model needs to "understand" human movement. It doesn't, because it never did.
"I have no evidence of your personal wealth. Buying 25,000 poses is not cheap. Therefore piracy." Read that sentence back to yourself slowly. You are publicly accusing a stranger of piracy because you do not know their bank balance. No evidence, no leaked file, no source name, nothing. You invented a financial profile for someone you have never met and concluded from it that they must have committed a crime. That is not common sense. That is libel with the words "common sense" stapled to the front of it.
The pattern across every one of your posts is the same. You speak with confidence about technical subjects you have not taken the time to understand. You ignore corrections when they come, then repeat the same incorrect claim two posts later as if the correction never happened. You refuse to look at the actual product, refuse to ask for the script that would settle the question, and use the resulting void of information as a license to invent the worst possible interpretation. That is not skepticism. That is not common sense. It is a posture, and it is wearing thin.
For the third time now. The script never produced 25,000 anatomically correct human poses. It produced 25,000 mathematical permutations of joint rotations on a blank Genesis 8 figure. Right arm raised 30 degrees, then 60, then 90. Left knee bent 45 degrees, combined with shoulder twisted 20 degrees. And so on, programmatically, through the rotation space of the rig. Most of those poses look weird. Many of them look impossible. That is the entire point.
You just wrote, word for word: "We know it is impossible for someone to produce 25,000 poses that could properly teach an AI how to pose a character." Correct. It would be impossible, and nobody tried. What I built is not a model that learned "what human poses look like." It is a model that learned how the Daz rig responds to inputs. A mapping table between a reference image and the rotation dials. The anatomy comes from the photo you feed it at inference time, not from the training data. The training set does not need to contain a single realistic human pose for the model to work, and it doesn't.
I told you this in my very first reply in this thread. I told you again in my last reply, in plain English: "It doesn't need to be trained on perfect, anatomically correct human poses. Why? Because it isn't generating poses out of thin air." You ignored it both times and went straight back to "impossible to produce 25,000 real poses," which is an argument against a tool I never described and never built. If you keep swinging at that strawman, that is on you, not on me.
And while we are on the subject of things you have ignored. I have offered the .dsa script that built the entire dataset at least three times now, openly, in this thread. Anyone who wants to see exactly how the data was generated can DM me and I will send it within the hour. Not you specifically, anyone. You haven't asked. None of the people making the loudest accusations in this thread have asked. You would rather speculate publicly than spend 60 seconds verifying the thing you are speculating about. That tells the audience here everything they need to know about whether this is a good-faith conversation.
While we are here, the three new accusations you packed into that post.
"High-definition renders." Each sample in the dataset is one .duf file plus seven viewport snapshots, 969x729 PNG, around 110 KB each. Those are the file properties of viewport screenshots, not ray-traced renders. A real HD render at that resolution, with lighting and materials, would be many times larger and take hundreds of times longer to produce. I can post the file properties of any sample on request. You won't ask, because verifying the claim would end the line of attack.
"Train your LLM." This is not an LLM. I told you that explicitly in my last reply and explained what it actually is, an image-conditioned regression head that maps a reference photo to joint-rotation values. You repeated the wrong term anyway. At this point that is a rhetorical choice, not a misunderstanding, and I think you know it. Calling it an LLM lets you keep arguing as if the model needs to "understand" human movement. It doesn't, because it never did.
"I have no evidence of your personal wealth. Buying 25,000 poses is not cheap. Therefore piracy." Read that sentence back to yourself slowly. You are publicly accusing a stranger of piracy because you do not know their bank balance. No evidence, no leaked file, no source name, nothing. You invented a financial profile for someone you have never met and concluded from it that they must have committed a crime. That is not common sense. That is libel with the words "common sense" stapled to the front of it.
The pattern across every one of your posts is the same. You speak with confidence about technical subjects you have not taken the time to understand. You ignore corrections when they come, then repeat the same incorrect claim two posts later as if the correction never happened. You refuse to look at the actual product, refuse to ask for the script that would settle the question, and use the resulting void of information as a license to invent the worst possible interpretation. That is not skepticism. That is not common sense. It is a posture, and it is wearing thin.
diogenese19348
Karma: 420
Sat, May 16Luxe, you are getting your troll points in today aren't you? I can understand wanting to protect people's IP rights. This isn't about that. This is about you trolling. You are doing a masterful job of it. You can stop anytime you want.
He told you how he did it. He is willing to provide the proof. You are being a troll.
The product is to translate pictures into DAZ poses. That's it. Nobody is supplying any text to AI about what they want the pose to look like. They take a picture, they drop it in the program. It produces a 3D pose from it. That's it.
He told you how he did it. He is willing to provide the proof. You are being a troll.
The product is to translate pictures into DAZ poses. That's it. Nobody is supplying any text to AI about what they want the pose to look like. They take a picture, they drop it in the program. It produces a 3D pose from it. That's it.
Luxe Muse
Karma: 7,226
Sat, May 16"I'm done addressing you specifically."... Anyways "Trained on 25,000+ pose files"
Luxe Muse
Karma: 7,226
Sat, May 16Trying to logically debate is called trolling now... common sense === trolling. Love the logic.
I just saw that it was not text-to-pose (I guess that would be the term), which means it is even worse. Because the image provided has to be interpreted, which might be where the images from the poses were for. To train a LoRA on top of maybe Gemma or Qwen’s CLIP vision.
His words were: Trained on 25,000+ pose files. I guess he also needs to disclose that, in addition to the 25,000 files, he used 175,000 high-definition renders (and those are also his words: 7 high-definition renders per pose).
I just saw that it was not text-to-pose (I guess that would be the term), which means it is even worse. Because the image provided has to be interpreted, which might be where the images from the poses were for. To train a LoRA on top of maybe Gemma or Qwen’s CLIP vision.
His words were: Trained on 25,000+ pose files. I guess he also needs to disclose that, in addition to the 25,000 files, he used 175,000 high-definition renders (and those are also his words: 7 high-definition renders per pose).
diogenese19348
Karma: 420
Sat, May 16I'm done with you. I'm reporting the thread, moderation can sort this out. You have made some very serious unsupported allegations in it.
laurendev
Karma: 445
Sat, May 16You are not trying to have a logical debate. You have been making accusations since your very first post in this thread. You ignore every answer I give you, then come back two posts later and repeat the exact same wrong claim as if nobody said anything.
Case in point: in your reply just now, you wrote "175,000 high-definition renders." I have already told you, twice, that nothing in the dataset was rendered. I told you the file size (around 110 KB). I told you the resolution (969x729). I told you what the API actually did (viewport screenshot, not a render). You wrote "high-definition renders" again anyway. That is not a debate. That is you closing your eyes and repeating yourself until you feel like you've won an argument.
I am confident that anyone who has actually been reading this thread can see exactly what is happening here. I am done responding
Case in point: in your reply just now, you wrote "175,000 high-definition renders." I have already told you, twice, that nothing in the dataset was rendered. I told you the file size (around 110 KB). I told you the resolution (969x729). I told you what the API actually did (viewport screenshot, not a render). You wrote "high-definition renders" again anyway. That is not a debate. That is you closing your eyes and repeating yourself until you feel like you've won an argument.
I am confident that anyone who has actually been reading this thread can see exactly what is happening here. I am done responding
Luxe Muse
Karma: 7,226
Sat, May 16@laurendev: Could it be the answer given does not satisfy reasoning?
@diogenese19348 uh... okay?
@diogenese19348 uh... okay?
Luxe Muse
Karma: 7,226
Sat, May 16Since I'm already accused of trolling, and being reported, let's make it right:
"It mathematically rotated joints on a blank character right in the viewport and took **high-res screenshots from 7 different* camera angles."
"High-definition renders." Each sample in the dataset is one .duf file plus seven viewport snapshots, 969x729 PNG, around 110 KB each.
"It mathematically rotated joints on a blank character right in the viewport and took **high-res screenshots from 7 different* camera angles."
"High-definition renders." Each sample in the dataset is one .duf file plus seven viewport snapshots, 969x729 PNG, around 110 KB each.
I wrote a review and pressed submit but I don't know where it went, or where I would see it.
REPLY
! REPORT
diogenese19348
Karma: 420
Wed, May 13It takes a while for them to post sometimes a full day, I think somebody reviews them first. It shows up on the product page you were reviewing. Yours is there now.
It's great, it's not perfect ( but the vendor says so on the product page ) My opinion is that it can get you so close to a pose from a photo, that very little adjustment is required. I've found with pose packs generally that I have to do some tweaking anyway since my figures are rarely the default shape, and of course poses have to be made to suit some figure shape, so the default shape is usually assumed as the target.
If I'm trying to reproduce a pose from a photo by hand I struggle to get the pelvis at the right angle, and if you change that later in the posing process most of the work you did on legs and arms will probably need to be done again, so having this program get me 90% of the way there is a huge help.
If I'm trying to reproduce a pose from a photo by hand I struggle to get the pelvis at the right angle, and if you change that later in the posing process most of the work you did on legs and arms will probably need to be done again, so having this program get me 90% of the way there is a huge help.
REPLY
! REPORT
I should add that I had some initial difficulty with getting it running, and the seller quickly found the problem and solved it. Also I had some feedback and suggestions and the seller quickly responded with very good answers.
REPLY
! REPORT
Honestly, and I'll probably get grilled for it on here, but it doesn't really bother me in the least. As long as the vendor is honest about it and adds some regular Iray renders with it, assuming the product is for Daz.
Personally, I'm more story focused and enjoy kit bashing, messing with the lights, environmental, tone mapping and surface settings to get the look I'm after. Found I can get closer to the look I'm after with those than relying on my mediocre skills with postwork.
Overall, I think everyone has their own reasons for doing the 3D thing, and for the most part, on the hobbiest, individualist level, I'm fine with people using whatever tools they have at their disposal and are comfortable with to get the look they're after as long as they're honest about it.
Sorry, alittle above and beyond the ai in the store topic.
Personally, I'm more story focused and enjoy kit bashing, messing with the lights, environmental, tone mapping and surface settings to get the look I'm after. Found I can get closer to the look I'm after with those than relying on my mediocre skills with postwork.
Overall, I think everyone has their own reasons for doing the 3D thing, and for the most part, on the hobbiest, individualist level, I'm fine with people using whatever tools they have at their disposal and are comfortable with to get the look they're after as long as they're honest about it.
Sorry, alittle above and beyond the ai in the store topic.
REPLY
! REPORT
Smells more like a tool to me and in this case it would be on the user to not "steal"... Biggest gripe with generative "AI" is that it trains on stolen material, here the tool is searching for the person and mapping position data to a genesis figure.. the user is responsible for not stealing art by respecting the artistic rights of the original image's owner. The modern use of "AI" is ridiculously broad.
REPLY
! REPORT
davidrobbins81
Karma: 468
Fri, May 15"The modern use of "AI" is ridiculously broad"
You got that right! I think that's one of the main reasons conversations like this get so heated. There's a big difference between using AI as a tool and having AI spit stuff out.
You got that right! I think that's one of the main reasons conversations like this get so heated. There's a big difference between using AI as a tool and having AI spit stuff out.
Pretty much the only things I've seen with this tag are backgrounds, reference images, and static (non-poseable/non-morphing) figures. Or at least so far, anyways. I haven't seen an actual rigged figure yet, but I'm sure that's coming eventually.
REPLY
! REPORT
Im not talking about AI art, or AI promos, I'm talking about the 3D models themselves are made using AI generation. and then sold as a product in the store.
Big difference.
Big difference.
REPLY
! REPORT
dirtrider00
Karma: 21,683
Mon, May 11Not sure how far along the whole AI modeling thing is, but in the end as much as I hate to say it, if the vendors don't or aren't making what the consumers want, then I can see some people moving to at least some AI type props and such. As a rough example, I've needed a mid to high quality baseball glove for G8 for years. I've asked, hinted, prayed and yet no one wants to make one. For me, I'll likely break down and buy one made in blender then scale and fit it as best I can to G8. For some of the simpler stuff, I can see people deciding to use a prop or such made with AI if they can't find it made the traditional way.
Though thinking about it, if there's a AI modeling program out there that anyone can use, why would you buy it through a vendor? I understand it for traditional items but something made with AI? Why not just do it yourself?
Though thinking about it, if there's a AI modeling program out there that anyone can use, why would you buy it through a vendor? I understand it for traditional items but something made with AI? Why not just do it yourself?
Tenserknot
Karma: 9,368
Tue, May 12absolutely, a customer can do it yourself. 25 words and you get what you want.
the point is, some vendors are turning around and selling that no work output as products. and both rendo and here dont see any issues with that.
and maybe there isn't a issue. except for some there is.
I think it reduced the usefulness of marketplaces. for the price of one AI generated model, a person can by a monthly subscription and generate 500 of them.
Guess thats the future.
the point is, some vendors are turning around and selling that no work output as products. and both rendo and here dont see any issues with that.
and maybe there isn't a issue. except for some there is.
I think it reduced the usefulness of marketplaces. for the price of one AI generated model, a person can by a monthly subscription and generate 500 of them.
Guess thats the future.
dirtrider00
Karma: 21,683
Tue, May 12I personally don't think, as of now AI is hurting the marketplaces, other than the "No AI" crowd. Basically most of it just feels over saturated. I'm basically only use G8.1 and G9 has several years under its belt already. Only so many ways you can cut a bikini and such. I tend to look at the texturing (colors and patterns) more than the item alot of the times. Kinda wish Foxy3d and BadKitty3d were a little more active. They tend to be more creative with their texturing/colors and such.
I know if I had the talent to model and texture stuff I'd be looking for the things that aren't in the stores. Oops, sorry, kinda getting off topic.
I know if I had the talent to model and texture stuff I'd be looking for the things that aren't in the stores. Oops, sorry, kinda getting off topic.
That's right, AI is coming, I'm using it, but there's still a big difference between the result produced by AI from a photo and the original photo. For example,
the photo of the sign
et le résultat du rendu avec Daz et Iray
the photo of the sign
et le résultat du rendu avec Daz et Iray
REPLY
! REPORT
Tenserknot
Karma: 9,368
Tue, May 12YOur looking at the bread, but the structure of the support pipes looks quite good.
bonj
Karma: 14,817
Tue, May 12Forgive me if I come across a little rude but I must speak my mind.
I find your products are very pore quality, they do not look real or very good at all. Do they even have all the correct texture maps? have you set up a displacement map? they look shiny and plastic.
Have you made any sales?
Did you imagine you could have a model shop online by only using your ai products?
How do you think this effects the stores of real 3d model makers?
As an individual I hold no bias towards you personally, however if every Tom, Dick and Harry also set up an ai model shop here don't you think it will eventually effect everyone else?
I get asked all the time now if my work is ai and I have lost sales because people assume I use AI when I absolutely do not , for anything.
Anyway I have said my piece, If you wish to grow and develop in 3d then learn some software.
I find your products are very pore quality, they do not look real or very good at all. Do they even have all the correct texture maps? have you set up a displacement map? they look shiny and plastic.
Have you made any sales?
Did you imagine you could have a model shop online by only using your ai products?
How do you think this effects the stores of real 3d model makers?
As an individual I hold no bias towards you personally, however if every Tom, Dick and Harry also set up an ai model shop here don't you think it will eventually effect everyone else?
I get asked all the time now if my work is ai and I have lost sales because people assume I use AI when I absolutely do not , for anything.
Anyway I have said my piece, If you wish to grow and develop in 3d then learn some software.
Tenserknot
Karma: 9,368
Tue, May 12I'm downloading your old five pack now to have a closer look, bonj, they look interesting and original.
bonj
Karma: 14,817
Tue, May 12@Tenserknot Thanks, they are very old now and my workflow is completely different. Some new freebies are coming soon showing my updated skills 
QUOTE: I'm talking about the 3D models themselves are made using AI generation. and then sold as a product in the store.
Big difference."
In the case of AI generated 3D models This sounds to me like a problem that solves itself.
I mean, considering the polarization on this issue. You're going to only have a few possible outcomes.
The hardcore anti AI activists will see that your model is marked "AI generated" and they won't buy it, nor will they bother generating their own models because that would be "haram."
Someone who is open to the use of AI ? well, they're not going to buy your model either. Because why would they, when they can just go to a site like meshy.ai and generate their own.
As much as I have embraced AI
I see the reality is this.
This permanent loss of scarcity, in the visual media production markets, cuts both ways
meaning that the sudden abundance of instant visual media production ,While, it may destroy the commercial freelance and even job markets for those who don't use the technology, it also completely devalues the output of those of us who are eagerly embracing the technology no matter how good it looks.
Because once people realize that great promotional ad that you created or Video or 3D model was made by using AI. They're going to immediately ask the. question. "Well, why should I pay you to do what I can go do for myself."
And I am perfectly okay with the economics of it being that I am newly retired ,with plenty of disposable income, and
And while yes, I'm still receiving small payments from the sales of my first AI assistant graphic novel through Kindle direct publishing.
I have no delusions about getting" rich" selling AI Generated output of any kind.
We are in the" attention economy " now,
YOUR ATTENTION needs to captured long enough for you to be exposed to to actual product idea or service.
All of the people who are heavily pro AI (that I follow-on Twitter/X )Spend most of their time obsessing over how many views and followers they get because Twitter/X gives payouts for engagement similar to what YouTube does, but on a smaller scale at this point.
Sure, the big tech companies that are selling the AI services are receiving HUGE. income from us users both. individuals and as corporate users.
But frankly in the end for most of us,
AI, it's just a tool to quickly produce and accelerate visual media for you to draw attention to something else that you're trying to sell or some other idea that you're trying to promote.
But the intrinsic value( such as it may be) better be in that" other actual thing" that you're selling or that "other actual idea" that you're promoting Because the visual media used to promote it has no value and is DISPOSABLE because everyone will be able to literally direct the production of good visual media. So once again, it will fall back to competition between ideas and narratives, not, who is "pouring their soul" into the making of some pretty pictures.
Big difference."
In the case of AI generated 3D models This sounds to me like a problem that solves itself.
I mean, considering the polarization on this issue. You're going to only have a few possible outcomes.
The hardcore anti AI activists will see that your model is marked "AI generated" and they won't buy it, nor will they bother generating their own models because that would be "haram."
Someone who is open to the use of AI ? well, they're not going to buy your model either. Because why would they, when they can just go to a site like meshy.ai and generate their own.
As much as I have embraced AI
I see the reality is this.
This permanent loss of scarcity, in the visual media production markets, cuts both ways
meaning that the sudden abundance of instant visual media production ,While, it may destroy the commercial freelance and even job markets for those who don't use the technology, it also completely devalues the output of those of us who are eagerly embracing the technology no matter how good it looks.
Because once people realize that great promotional ad that you created or Video or 3D model was made by using AI. They're going to immediately ask the. question. "Well, why should I pay you to do what I can go do for myself."
And I am perfectly okay with the economics of it being that I am newly retired ,with plenty of disposable income, and
And while yes, I'm still receiving small payments from the sales of my first AI assistant graphic novel through Kindle direct publishing.
I have no delusions about getting" rich" selling AI Generated output of any kind.
We are in the" attention economy " now,
YOUR ATTENTION needs to captured long enough for you to be exposed to to actual product idea or service.
All of the people who are heavily pro AI (that I follow-on Twitter/X )Spend most of their time obsessing over how many views and followers they get because Twitter/X gives payouts for engagement similar to what YouTube does, but on a smaller scale at this point.
Sure, the big tech companies that are selling the AI services are receiving HUGE. income from us users both. individuals and as corporate users.
But frankly in the end for most of us,
AI, it's just a tool to quickly produce and accelerate visual media for you to draw attention to something else that you're trying to sell or some other idea that you're trying to promote.
But the intrinsic value( such as it may be) better be in that" other actual thing" that you're selling or that "other actual idea" that you're promoting Because the visual media used to promote it has no value and is DISPOSABLE because everyone will be able to literally direct the production of good visual media. So once again, it will fall back to competition between ideas and narratives, not, who is "pouring their soul" into the making of some pretty pictures.
REPLY
! REPORT
Tenserknot
Karma: 9,368
Tue, May 12This is the crux of it. the tool allows the customer to be their own supplier.
I think it's a good initiative and worth appreciating compared to having nothing stated at all, but I've stopped buying anything here or on Renderosity.
The reason is that some creators use AI without disclosing it. I've reported it several times, but in some cases nothing improves, and it has become unclear which works involve AI use.
I'm probably in the minority, and I don't think my behavior will have much impact, but in the future it might lead to reduced sales even for creators who don't use AI.
The reason is that some creators use AI without disclosing it. I've reported it several times, but in some cases nothing improves, and it has become unclear which works involve AI use.
I'm probably in the minority, and I don't think my behavior will have much impact, but in the future it might lead to reduced sales even for creators who don't use AI.
REPLY
! REPORT
bonj
Karma: 14,817
Tue, May 12I feel exactly the same, stopped buying from here , rendo, shards and any store actively promoting ai over reel artist.
I only buy models I really want and if it supports a particular artist / model maker.
I can make most things myself but do like to collect and use other peoples work, I like to interact with them and get to know the person behind the product. I have made a lot of friends I have great respect for and genuinely care about them and their's. I believe this is called community in an industry and it builds a strong foundation. AI destroys this. Simply the community breaks down , conflicts and arguments develop. Trust gets lost and we end up like this. I won't buy very much on any platform now so to all vendors( I hate that word- sounds like a machine). To all vendors please provide a link to a personal site or a paypal alternative for a private sale.
I honeltly don't know what else to suggest, we have been forced into this position. I'd rather not share my models then promote ai on any site.
We the model maker are the product here. Our patience and skill developed over many years and at much cost for the software and hardware. Why should we share with AI?
I only buy models I really want and if it supports a particular artist / model maker.
I can make most things myself but do like to collect and use other peoples work, I like to interact with them and get to know the person behind the product. I have made a lot of friends I have great respect for and genuinely care about them and their's. I believe this is called community in an industry and it builds a strong foundation. AI destroys this. Simply the community breaks down , conflicts and arguments develop. Trust gets lost and we end up like this. I won't buy very much on any platform now so to all vendors( I hate that word- sounds like a machine). To all vendors please provide a link to a personal site or a paypal alternative for a private sale.
I honeltly don't know what else to suggest, we have been forced into this position. I'd rather not share my models then promote ai on any site.
We the model maker are the product here. Our patience and skill developed over many years and at much cost for the software and hardware. Why should we share with AI?
Tenserknot
Karma: 9,368
Tue, May 12This was supposed to be the site that favored artists over AI, that's why its so unexpected.
Everwild
Karma: 8,978
Wed, May 13I never expected to like using Daz. I tried it and it just sort of stuck. It has always felt a bit like collaborative art to me, because I take the beautiful models others make and I turn them into the characters I've had in my head for years. I guess I'm grateful for the few years I had before all of this to learn lighting and make better images. Maybe with another few years, I could even stretch further beyond the limited capabilities I currently possess.
3D printing made things easier for people, too. I think the difference is that no one is shoving a 3D printer down my throat and making it impossible for me to sculpt without being able to opt out of that technology (and personally, this is one of the things I resent most about the onset of AI).
Now, I've grown distrustful of the Daz store, too. And, the thing that I really just cannot wrap my head around is why anyone would so willingly embrace their own obsolescence:
"Buy my AI-slop!"
Um, why? AI has made everything so accessible, we can now just generate our own slop, kthxbai!
"Wait. Why can't I make my house payment this month?"
(Not at all intended to make light of those vendors/artists with genuine concerns about their future, because I expect you're out there, too.)
3D printing made things easier for people, too. I think the difference is that no one is shoving a 3D printer down my throat and making it impossible for me to sculpt without being able to opt out of that technology (and personally, this is one of the things I resent most about the onset of AI).
Now, I've grown distrustful of the Daz store, too. And, the thing that I really just cannot wrap my head around is why anyone would so willingly embrace their own obsolescence:
"Buy my AI-slop!"
Um, why? AI has made everything so accessible, we can now just generate our own slop, kthxbai!
"Wait. Why can't I make my house payment this month?"
(Not at all intended to make light of those vendors/artists with genuine concerns about their future, because I expect you're out there, too.)
davidrobbins81
Karma: 468
Fri, May 15"any store actively promoting ai over reel artist"
I don't see Renderhub actively promoting AI. I see them giving users tools to actively hide it if they choose.
I don't see Renderhub actively promoting AI. I see them giving users tools to actively hide it if they choose.
QUOTE: "This is the crux of it. the tool allows the customer to be their own supplier."
It's what they call " vertical integration " in the business world.
People don't seem to have a problem when big companies do it, such as when Apple became their own GPU supplier reducing external dependencies on others, which turned out to be quite the wise move Considering all the windows PC users now panicking that NVIDIA appears to have abandoned the retail consumer GPU market in favor of enterprise solutions for data centers.
But when a single individual decides to vertically integrate the creation of his images, animations,3D models and even his marketing materials into his own internal pipeline ,using AI,
All of a sudden, people, who specialized in those single areas, begin behaving as though they have some fundamental international human right to be hired by you to do those tasks instead.
This attitude is particularly prevalent in the indie comic space on Twitter/X
In a recent thread, someone, (presumably white) was promoting his latest indie comic when someone else complained that all of the characters in his comic appeared to be white and there was no so-called "black representation."
Now I ,A black man myself, posted a link to my recent graphic novel (80 pages produced with AI in 22 days.), which prominently features one of the main protagonists ,on my superhero team, who happens to be black.
I bluntly suggest they stop harassing the white comic creator and consider using use AI tools to create their own Black Comic characters
I was immediately dog piled and called all sorts of vile names with death threats from the two self, identified black people in the thread.
These very same black people ,who moments earlier, were admonishing a white comics writer for not providing them with black characters in his story,
began viciously assailing a fellow black person for having the audacity to create my own story and my own black characters using AI, instead of hiring black artist and writers and spending money on a book that will never turn a profit as I'm doing this just as a retired hobbyist.
I will be getting some physical prints made at my own expense, just for my own personal collection and distribution to family and close friends who might be interested.
This is what seems to upset many people the most is the freedom we are given from old legacy permission structures to which they remain willingly enslaved.
It's what they call " vertical integration " in the business world.
People don't seem to have a problem when big companies do it, such as when Apple became their own GPU supplier reducing external dependencies on others, which turned out to be quite the wise move Considering all the windows PC users now panicking that NVIDIA appears to have abandoned the retail consumer GPU market in favor of enterprise solutions for data centers.
But when a single individual decides to vertically integrate the creation of his images, animations,3D models and even his marketing materials into his own internal pipeline ,using AI,
All of a sudden, people, who specialized in those single areas, begin behaving as though they have some fundamental international human right to be hired by you to do those tasks instead.
This attitude is particularly prevalent in the indie comic space on Twitter/X
In a recent thread, someone, (presumably white) was promoting his latest indie comic when someone else complained that all of the characters in his comic appeared to be white and there was no so-called "black representation."
Now I ,A black man myself, posted a link to my recent graphic novel (80 pages produced with AI in 22 days.), which prominently features one of the main protagonists ,on my superhero team, who happens to be black.
I bluntly suggest they stop harassing the white comic creator and consider using use AI tools to create their own Black Comic characters
I was immediately dog piled and called all sorts of vile names with death threats from the two self, identified black people in the thread.
These very same black people ,who moments earlier, were admonishing a white comics writer for not providing them with black characters in his story,
began viciously assailing a fellow black person for having the audacity to create my own story and my own black characters using AI, instead of hiring black artist and writers and spending money on a book that will never turn a profit as I'm doing this just as a retired hobbyist.
I will be getting some physical prints made at my own expense, just for my own personal collection and distribution to family and close friends who might be interested.
This is what seems to upset many people the most is the freedom we are given from old legacy permission structures to which they remain willingly enslaved.
REPLY
! REPORT
Tenserknot
Karma: 9,368
Tue, May 12The biggest problem which is ongoing, is the ability to think independently. lot of sheeple follow the crowd.
At this point AI in entrenched and growing. None of us can change that. Venture capital has already decided it is the future.
No amount of screaming at the approaching tidal wave is going to stop it in any way.
Good luck with your comicbook and your vertically integrated studio. AI agents are going to take away a lot of jobs, might as well put one to work for you.
At this point AI in entrenched and growing. None of us can change that. Venture capital has already decided it is the future.
No amount of screaming at the approaching tidal wave is going to stop it in any way.
Good luck with your comicbook and your vertically integrated studio. AI agents are going to take away a lot of jobs, might as well put one to work for you.
Several posts in this thread were reported by community members and have been removed for violating the Terms of Use.
We want to remind everyone to review the User Conduct section of the RenderHub Terms of Use. Users will not use the website for any activity that affects the normal operations of the community. Personal attacks, harassment, and attempts to incite conflict are not acceptable here.
Our goal is to keep RenderHub a place for creativity, support, and civil conversation. We appreciate your cooperation in helping to maintain that standard.
We want to remind everyone to review the User Conduct section of the RenderHub Terms of Use. Users will not use the website for any activity that affects the normal operations of the community. Personal attacks, harassment, and attempts to incite conflict are not acceptable here.
Our goal is to keep RenderHub a place for creativity, support, and civil conversation. We appreciate your cooperation in helping to maintain that standard.
REPLY
! REPORT
bonj
Karma: 14,817
Wed, May 13Fantastic, how about deleting some of the ai slop that get repeatedly posted across many threads, cheers.
RenderHub Admin 1
Admin: 15,980
Wed, May 13Per our AI Content Policy and Guidelines:
"...AI-generated images in forum posts are allowed to be used to continue a conversation, to illustrate a concept or support a point of view..."
https://www.renderhub.com/info/ai-content-policy
"...AI-generated images in forum posts are allowed to be used to continue a conversation, to illustrate a concept or support a point of view..."
https://www.renderhub.com/info/ai-content-policy
bonj
Karma: 14,817
Wed, May 13What is the policy around marketing ai aggressively and promoting a youtube slop accounts all at the same time as insulting industry professionals? It makes the professionals leave the site, would this be renderhubs preference?
To leave that low effort ai cartoon here (above) is to reward his actions. are you aware of that? Look at it, 3 low effort characters that even all have the exact same pose. You want that slop over real 3d content? If that is the case then I will go elsewhere but I won't go quietly.
To leave that low effort ai cartoon here (above) is to reward his actions. are you aware of that? Look at it, 3 low effort characters that even all have the exact same pose. You want that slop over real 3d content? If that is the case then I will go elsewhere but I won't go quietly.
RenderHub Admin 1
Admin: 15,980
Wed, May 13At RenderHub, we value free expression and encourage a diversity of thought.
Except in extreme cases, we only remove forum posts that directly violate the Terms of Use and/or published policies.
Our goal is to ensure that the community remains a place where all voices can be heard, as long as they are shared in a constructive and civil way.
Except in extreme cases, we only remove forum posts that directly violate the Terms of Use and/or published policies.
Our goal is to ensure that the community remains a place where all voices can be heard, as long as they are shared in a constructive and civil way.
bonj
Karma: 14,817
Fri, May 15@RenderHub Admin 1
You still have not addressed the bigger issue here. A certain individual who seems to have a hobby in insulting your site users while pushing a youtube channel. Why else would he immediately post another" the film industry is dead" comment if it were not to cause more conflict.
Not to mention the comment bellow also designed to get a rise, Your idea of freedom of speech is very one sided and it encourages conflict.
You still have not addressed the bigger issue here. A certain individual who seems to have a hobby in insulting your site users while pushing a youtube channel. Why else would he immediately post another" the film industry is dead" comment if it were not to cause more conflict.
Not to mention the comment bellow also designed to get a rise, Your idea of freedom of speech is very one sided and it encourages conflict.
I think the problem here is not the AI policy per se. Is someone diverting the narrative y making excuses to continuously post his videos to promote his channel.
There was no need to place that video there, the anecdote about 'black representation' don't even had anything to do with AI except the poster fixation to make everything about it.
There was no need to place that video there, the anecdote about 'black representation' don't even had anything to do with AI except the poster fixation to make everything about it.
REPLY
! REPORT
Bobb
Karma: 1,412
Thu, May 14We could always move these conversations over to Sade's forums at 3DShards. Her place is (thankfully) AI slop free.
bonj
Karma: 14,817
Thu, May 14Good idea, I just noticed as of April 13, 2026, 3DShards will no longer accept new products that have utilized AI in their creation process.
This is fantastic news for people like myself.
This is fantastic news for people like myself.
stevenjoseph8844
Karma: 2,577
Thu, May 143D Shards? That website sucks. It might be "AI slop free", but the site looks like it's just "slopped" together. It's a total mishmash of WordPress plugins. I'm surprised that any quality seller would want their products to appear there.
bonj
Karma: 14,817
Fri, May 15If shards are going to stand by the artist then I will stand with them.
Quote:At RenderHub, we value free expression and encourage a diversity of thought.
And for that we are thankful
long live Renderhub!!!
And for that we are thankful
long live Renderhub!!!
REPLY
! REPORT
stevenjoseph8844
Karma: 2,577
Thu, May 14Even though I don't like AI-generated images and videos, I'll have to agree with you on this. It wouldn't be right for RenderHub to censor everyone who had a differing opinion.
I'm not surprised there is so much negativity surrounding AI, it's a scary subject, but at the bottom of it, it is only another tool. In the hands of a mediocre artist it turns out slop, the same way 3D rendering, photo-editing, and cave drawings do.
The problem is it turns out massive amounts of slop quickly, but in the hands of a talented creative, it turns out some really impressive work..
It also has the capability of completely changing how we go about things, the same way, well Lotus 1-2-3 changed how accounting was done, WordPerfect changed document creation, Autocad changed drafting.... the point is we've been dealing with computer automation of tasks for decades now. This really isn't anything new.
Change is both exciting and scary. It's also pretty unstoppable, life is all about change.
So how is writing double entry ledger lines with a quill pen by the light of a candle coming along Bob Cratchit?
If our communications systems are allowing a flood of mediocre works, no matter how they were created, it's the fault of our communications systems, not AI. Facebook, TicTok, Youtube, Spottily, etc were flooded with slop long before AI added its pile.
If AI allows me to do something quicker and better than I am doing it now, I'm going to use it. If it doesn't, I'm not going to bother with it.
If an artist uses AI to create a great product I want to use, I'll buy it. If it's a pile of slop I won't, just like I won't by slop created with 3D tools.
Isn't that the intelligent way to go about it?
The problem is it turns out massive amounts of slop quickly, but in the hands of a talented creative, it turns out some really impressive work..
It also has the capability of completely changing how we go about things, the same way, well Lotus 1-2-3 changed how accounting was done, WordPerfect changed document creation, Autocad changed drafting.... the point is we've been dealing with computer automation of tasks for decades now. This really isn't anything new.
Change is both exciting and scary. It's also pretty unstoppable, life is all about change.
So how is writing double entry ledger lines with a quill pen by the light of a candle coming along Bob Cratchit?
If our communications systems are allowing a flood of mediocre works, no matter how they were created, it's the fault of our communications systems, not AI. Facebook, TicTok, Youtube, Spottily, etc were flooded with slop long before AI added its pile.
If AI allows me to do something quicker and better than I am doing it now, I'm going to use it. If it doesn't, I'm not going to bother with it.
If an artist uses AI to create a great product I want to use, I'll buy it. If it's a pile of slop I won't, just like I won't by slop created with 3D tools.
Isn't that the intelligent way to go about it?
REPLY
! REPORT
darkart1
Karma: 2,417
Fri, May 15"The problem is it turns out massive amounts of slop quickly"
"...etc were flooded with slop long before AI added its pile"
You're right about that. There was plenty of garbage long before AI was a thing. Now it's a garbage avalanche.
"...etc were flooded with slop long before AI added its pile"
You're right about that. There was plenty of garbage long before AI was a thing. Now it's a garbage avalanche.
bonj
Karma: 14,817
Fri, May 15ai is not just another tool. it's a financial weapon designed to destroy industries and force all resource into the hands of a small group of already rich old men. The 4th industrial revolution is not designed to aid civilization and give us the tools to produce anything, it's designed to take everything we the little people have, including our freedom and later our bodily autonomy. Make your ai assisted "art" all you like but sooner or later you will be forced to see the bigger picture.
diogenese19348
Karma: 420
Fri, May 15bonj, while huge corporations are putting up massive data centers for centralized processing there is nothing inherent in AI technology that dictates that needs to be the case. Yeah, Microsoft, Apple, Google et al would like you to be strapped to their vision, but you know, there is an independent programmer from Turkey who just harnessed AI to run in 4 GB of VRAM on your desktop to translate pictures into DAZ3D pose files. That isn't a fluke. Agents can run independently of the web and data centers and five years from now the concept of building huge data centers to run AI may seem like a quaint old fashioned idea.
It all depends on whether society allows a few huge corporations to monopolize things, but we have the same issues with things like medical care, drug companies, right to repair, and privacy. AI is just another front in the same battle.
It all depends on whether society allows a few huge corporations to monopolize things, but we have the same issues with things like medical care, drug companies, right to repair, and privacy. AI is just another front in the same battle.
bonj
Karma: 14,817
Fri, May 15Yes exactly it's all very invasive, overwhelming and extremely destructive on all fronts. That 4 g of vram will get lower I imagine.
Used to be you needed a 100 person crew to get cinematic footage.
Now a just a couple people can produce, film and edit, an entire feature film in a week.
Now a just a couple people can produce, film and edit, an entire feature film in a week.
REPLY
! REPORT
evarest
Karma: 220
Fri, May 15By that logic you can spend months in pastry school learning how to bake extraordinary cakes.
Or you can walk into a bathroom with a silver tray and call whatever you put on it a cake.
Technically, they’re both “cakes.If given the choice, which one would most people actually want to eat?
Or you can walk into a bathroom with a silver tray and call whatever you put on it a cake.
Technically, they’re both “cakes.If given the choice, which one would most people actually want to eat?
diogenese19348
Karma: 420
Fri, May 15That's not a particularly good analogy, the technology for film making has changed vastly, the technology for baking hasn't, although there have been advances in measuring and keeping a steady temperature in the oven which helps.
You still need the training to make a good movie, the tools to make it have just changed. Has for the music industry too, It costs much less to produce an album now then when the Beatles were doing it. Or even 30 years ago for that matter.
But without the proper training, it is just as easy to make a crappy movie today as it was back in the days of the Sci-Fi "B" movie. It just takes less time and expense
And lord knows it has gotten way quicker to make crappy music
You still need the training to make a good movie, the tools to make it have just changed. Has for the music industry too, It costs much less to produce an album now then when the Beatles were doing it. Or even 30 years ago for that matter.
But without the proper training, it is just as easy to make a crappy movie today as it was back in the days of the Sci-Fi "B" movie. It just takes less time and expense
And lord knows it has gotten way quicker to make crappy music
bonj
Karma: 14,817
Fri, May 15@diogenese19348
I completely agree with what you said here. However the real revolution has already happened with digital in the mid 90s. generative movies will and have already become a reality but few will ever be taken seriously for exactly the reason you stated.
30 years ago you could only include up to 5 mins of digital footage per 60 mins in your broadcast production. This soon changed but for around 5 years we low budget Indy's had to find workarounds or hire expensive pro kit taking a large chunk of our budgets.
The real revolution was digital editing as cutting film was very expensive. I shot on digital a lot back then and was one of the first nonlinear editors in my country. I still have some of my mini dv backup masters that were saved from a warehouse fire.
In all honesty I believe using ai to take shortcuts will dumb down the creative process, This is the underlying reason I personally will not use anything generative. I enjoy what I do and do not want to farm it out to the corporations.
I completely agree with what you said here. However the real revolution has already happened with digital in the mid 90s. generative movies will and have already become a reality but few will ever be taken seriously for exactly the reason you stated.
30 years ago you could only include up to 5 mins of digital footage per 60 mins in your broadcast production. This soon changed but for around 5 years we low budget Indy's had to find workarounds or hire expensive pro kit taking a large chunk of our budgets.
The real revolution was digital editing as cutting film was very expensive. I shot on digital a lot back then and was one of the first nonlinear editors in my country. I still have some of my mini dv backup masters that were saved from a warehouse fire.
In all honesty I believe using ai to take shortcuts will dumb down the creative process, This is the underlying reason I personally will not use anything generative. I enjoy what I do and do not want to farm it out to the corporations.
diogenese19348
Karma: 420
Fri, May 15If you've never heard of them, The Stupendium. They just bought a robot camera track to shoot this one with. There is one person in the video. All practical effects and editing.
I think that's true, and I think it's one of the reasons I dislike a lot of AI generated content. It's all 'look at the special effects', 'look at the perfect lighting', 'Look at the lipsync', 'look at the characters', well ok but if the story is crap then it's crap. Also the number of TV shows I've seen where the writer(s) had one good idea, and the rest of it is padded out with recycled stuff.
REPLY
! REPORT
davidrobbins81
Karma: 468
Fri, May 15"if the story is crap then it's crap"
True! AI can't replace the true creative mind. It just sucks that there's SO MUCH of this crap.
True! AI can't replace the true creative mind. It just sucks that there's SO MUCH of this crap.
For anyone who doe's not understand why I'm so passionate for my craft and trade I recommend this helpful and informative channel. There's much more to cinema than just the images.
The House of Tabula https://www.youtube.com/@TheHouseofTabula/videos
Art is Not Content
The House of Tabula https://www.youtube.com/@TheHouseofTabula/videos
Art is Not Content
REPLY
! REPORT
Luxe Muse
Karma: 7,226
Sat, May 16Thank you for sharing the channel. I am not sure why I had not come across it before. It is right up my alley.
Personal anecdote. I remember my better half suffering after being forced to train the person who would take over his job. He never complained, but we all felt it in the house. He perhaps had a choice not to, but he wanted to stretch our livelihood as best as he could while trying to find a new job. That was very bad of my better half's employer, and to this day I have banned their services from our lives.
But AI (humans using it of course,) is not giving artists that choice. Even dead artists are being forced to train their replacements.
Personal anecdote. I remember my better half suffering after being forced to train the person who would take over his job. He never complained, but we all felt it in the house. He perhaps had a choice not to, but he wanted to stretch our livelihood as best as he could while trying to find a new job. That was very bad of my better half's employer, and to this day I have banned their services from our lives.
But AI (humans using it of course,) is not giving artists that choice. Even dead artists are being forced to train their replacements.
Quote: Used to be you needed a 100 person crew to get cinematic footage.
Now a just a couple people can produce, film and edit, an entire feature film in a week.
Welcome to the age of post, scarcity in the creative fields..
It was the natural organic scarcity that existed amongst the general population, whereby only a small handful of people, Where actually capable of producing visual media to say nothing of really good "art."
it is my opinion that at some point, people began to attach their entire identities to their artistic processes rather than the outcome of those processes.
This enabled people to insulate themselves in myopic little mutual validation communities where everyone was equally celebrated as long as they performed the correct process in creating their "art."
No matter how objectively good or bad the outcome.
This feeling of self-importance seems particularly exacerbated if the approved processes is slow painful and tedious
because it resulted in actual "suffering" on the part of the artist and they believe that this suffering somehow added more humanity or " soul" to the final result of their process.
Add to that in the modern world, these processes
demand that you go in hire fellow artists, who specialized in adjacent areas that you don't, as some sort of moral obligation.
I had a moderator at blender artist.org. call me "cheap" because I used a throw away AI Generated jingle for some cartoon animations that I made for my young grandsons
made with the REALLUSION cartoon, animator software.
(not AI)
Now a just a couple people can produce, film and edit, an entire feature film in a week.
Welcome to the age of post, scarcity in the creative fields..
It was the natural organic scarcity that existed amongst the general population, whereby only a small handful of people, Where actually capable of producing visual media to say nothing of really good "art."
it is my opinion that at some point, people began to attach their entire identities to their artistic processes rather than the outcome of those processes.
This enabled people to insulate themselves in myopic little mutual validation communities where everyone was equally celebrated as long as they performed the correct process in creating their "art."
No matter how objectively good or bad the outcome.
This feeling of self-importance seems particularly exacerbated if the approved processes is slow painful and tedious
because it resulted in actual "suffering" on the part of the artist and they believe that this suffering somehow added more humanity or " soul" to the final result of their process.
Add to that in the modern world, these processes
demand that you go in hire fellow artists, who specialized in adjacent areas that you don't, as some sort of moral obligation.
I had a moderator at blender artist.org. call me "cheap" because I used a throw away AI Generated jingle for some cartoon animations that I made for my young grandsons
made with the REALLUSION cartoon, animator software.
(not AI)
REPLY
! REPORT
"made it impossible for small independent creators to even get a project off the ground unless he submits to these top down permission regimes."
What a ridiculous concept this is. Digital has been here 30 years. Everyone now has a camera in their pocket and pro video editing software is available for free. Distribution has never been so easy. There are no restrictions and no establishment to conform to.
All it take to make your own production now is the effort and motivation.
What a ridiculous concept this is. Digital has been here 30 years. Everyone now has a camera in their pocket and pro video editing software is available for free. Distribution has never been so easy. There are no restrictions and no establishment to conform to.
All it take to make your own production now is the effort and motivation.
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15My company had 3 full time directors in a partnership. An external crew of around 5 other key people. We did all the filming , editing, script writing, soundtrack, coding, absolutely all in house.
davidrobbins81
Karma: 468
Fri, May 15"Digital has been here 30 years"
So let me get this straight, because you sound a bit hypocritical in this thread. What side of the fence were you on when digital was replacing film? And how long did it take for you to finally accept reality?
So let me get this straight, because you sound a bit hypocritical in this thread. What side of the fence were you on when digital was replacing film? And how long did it take for you to finally accept reality?
bonj
Karma: 14,817
Fri, May 15I was young then and ambitious. I adopted digital very early for my own productions but for freelance work film was still top dog.
Digital made production cheap, especially editing. An avid desk and later on pc with premier or speed razor made it accessible.
Traditional production rules still applied especially if wanting to find employment and digital was a great way to get experience.
I don't care what tools people use but I care what I consume and if any craft went into it. You could make a very good ai movie or art movie if it was done with the right mindset. I likely won't consume it but good luck...
I'm in a position were I can stay traditional.
Ai is not easy to compare to the digital revolution, they are very different beast. The real issue here is that ai is being pushed on us and when we refuse to eat it! we get attacked.
Anyway I'm off for the weekend but feel free to ask anything else.
Digital made production cheap, especially editing. An avid desk and later on pc with premier or speed razor made it accessible.
Traditional production rules still applied especially if wanting to find employment and digital was a great way to get experience.
I don't care what tools people use but I care what I consume and if any craft went into it. You could make a very good ai movie or art movie if it was done with the right mindset. I likely won't consume it but good luck...
I'm in a position were I can stay traditional.
Ai is not easy to compare to the digital revolution, they are very different beast. The real issue here is that ai is being pushed on us and when we refuse to eat it! we get attacked.
Anyway I'm off for the weekend but feel free to ask anything else.
Omg everybody watch the slop I generated this morning omg I destroyed cinema industry hollywood is sueing me but you will never understand that 3D is deadh. Daz 3D is dead Genesis 9>Victoria 5.1
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15The other vid you posted was an entertaining 40 mins thanks for the laughs.
Quote: What a ridiculous concept this is. Digital has been here 30 years. Everyone now has a camera in their pocket and pro video editing software is available for free.
You mean same digital video tech that DESTROYED the careers of the traditional chemical film engineers
at Eastman kodak??
Quote:Distribution has never been so easy. There are no restrictions and no establishment to conform to
Really?? try getting a "distribution deal" of a film
that is critical of certain… (unnamed groups or nation states)
or try getting a Payment processor to process income from online supporters if your movie/prints for sale etc. are declared to be "problematic"
Kickstarter just banned "adult" content.( no more crown funding of nudie girl comics)
Quote: My company had 3 full time directors in a partnership. An external crew of around 5 other key people. We did all the filming , editing, script writing, soundtrack, coding, absolutely all in house.'
Lovely Mate, what types of films are you making?
I like Action ,intrigue VXF heavy,hard sci-fi/space epics,(with good stories of course)
Shooting it live on location is tens of thousands of dollars in permits alone before the cost of insurance for the entire production.
and sorry "phone cameras" just won't cut it for what I may want so there is pro equipment rentals and catering along side paying human actors .
But silly me, I can just use Unreal engine and the free
metahumans and do it all in house with "traditional" CG right?
Sure I can try but if I want some really professional
animation, at volume I need to hire a team of professionals
Pay 10K just to rent the studio for performance capture
and hire VFX pros who are expert in Houdini & Maya
and spend the next three or four years paying ongoing funding myproject.
or I could Just do what this former Unreal Engine filmmaker did,
and spend about $400 USD in "generation credits" and spend 4 days to produce films like this.
You mean same digital video tech that DESTROYED the careers of the traditional chemical film engineers
at Eastman kodak??
Quote:Distribution has never been so easy. There are no restrictions and no establishment to conform to
Really?? try getting a "distribution deal" of a film
that is critical of certain… (unnamed groups or nation states)
or try getting a Payment processor to process income from online supporters if your movie/prints for sale etc. are declared to be "problematic"
Kickstarter just banned "adult" content.( no more crown funding of nudie girl comics)
Quote: My company had 3 full time directors in a partnership. An external crew of around 5 other key people. We did all the filming , editing, script writing, soundtrack, coding, absolutely all in house.'
Lovely Mate, what types of films are you making?
I like Action ,intrigue VXF heavy,hard sci-fi/space epics,(with good stories of course)
Shooting it live on location is tens of thousands of dollars in permits alone before the cost of insurance for the entire production.
and sorry "phone cameras" just won't cut it for what I may want so there is pro equipment rentals and catering along side paying human actors .
But silly me, I can just use Unreal engine and the free
metahumans and do it all in house with "traditional" CG right?
Sure I can try but if I want some really professional
animation, at volume I need to hire a team of professionals
Pay 10K just to rent the studio for performance capture
and hire VFX pros who are expert in Houdini & Maya
and spend the next three or four years paying ongoing funding myproject.
or I could Just do what this former Unreal Engine filmmaker did,
and spend about $400 USD in "generation credits" and spend 4 days to produce films like this.
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15Please before responding take the time to watch some of The House of Tabula video's. I'm sure you would agree with much of it especially the parts about "outsider art" referencing many of the aspects you would resonate with. This argument should at least come to an end with us both disagreeing and that is fine. But be more open to others and maybe your opinions would be taken more seriously.
A camera will cost from around 500 quid a week. Hire it with a monitor and jib maybe even a tracking system and Steadicam and get an amazing deal. For low budget productions film off season as the hire house will have loads of kit not being used and will be happy to offer a deal.
If you do your pre production and schedule correctly you can shoot very low budget and take your time in post to add anything you like with effects. Ideally if you are a qualified or experienced animator and even model maker then you don't need a large fx crew as you can do it all yourself.
https://www.gearbooker.com/en/rent-cinema-video-cameras-sony-fx9-in-exeter-42263-l?SearchText=Sony+PXW-FX9+pa
https://vmi.tv/product-category/camera-hire/page/2/
If you do your pre production and schedule correctly you can shoot very low budget and take your time in post to add anything you like with effects. Ideally if you are a qualified or experienced animator and even model maker then you don't need a large fx crew as you can do it all yourself.
https://www.gearbooker.com/en/rent-cinema-video-cameras-sony-fx9-in-exeter-42263-l?SearchText=Sony+PXW-FX9+pa
https://vmi.tv/product-category/camera-hire/page/2/
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15I think someone is just frustrated and angry because they didn't go for it when they were young and now it's too late.
I'm sorry Tenserknot if this thread has fallen off track a couple times. I'm going to leave it for now and go do some work. Best regards everybody.
REPLY
! REPORT
Quote:I think someone is just frustrated and angry because they didn't go for it when they were young and now it's too late.
When I was young I used the tools of the time
as professional Graphic Designer for print
when any new tech was invented I embraced it which kept me employed enough to raise a family .
Now I am retired and just in time to enjoy creating graphic novels & short animations with incredible new AI tools.
Quote":Please before responding take the time to watch some of The House of Tabula video's. I'm sure you would agree with much of it especially the parts about "outsider art" referencing many of the aspects you would resonate with"
Saw it and…no ...just a bunch of the usual lofty self validating platitudes that justify strict adherence to the existing permission structures replete with nebulous definitions of "cannonized" art that have dictated HOW are visual media must be created and interpreted by the "enlightened."
I would take these online/youtube "art experts" more seriously if so many had not exposed themselves as complete phonies by viciously trashing an actual Monet painting ,like preprogrammed NPC's ,after hearing the term "AI" associated with it.
they could not even be bothered to do a reverse google image search before embarrassing themselves, they heard "AI" and went into rage filled attack mode.
Oh no Monet is slop!!
When I was young I used the tools of the time
as professional Graphic Designer for print
when any new tech was invented I embraced it which kept me employed enough to raise a family .
Now I am retired and just in time to enjoy creating graphic novels & short animations with incredible new AI tools.
Quote":Please before responding take the time to watch some of The House of Tabula video's. I'm sure you would agree with much of it especially the parts about "outsider art" referencing many of the aspects you would resonate with"
Saw it and…no ...just a bunch of the usual lofty self validating platitudes that justify strict adherence to the existing permission structures replete with nebulous definitions of "cannonized" art that have dictated HOW are visual media must be created and interpreted by the "enlightened."
I would take these online/youtube "art experts" more seriously if so many had not exposed themselves as complete phonies by viciously trashing an actual Monet painting ,like preprogrammed NPC's ,after hearing the term "AI" associated with it.
they could not even be bothered to do a reverse google image search before embarrassing themselves, they heard "AI" and went into rage filled attack mode.
Oh no Monet is slop!!
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15ART IS NOT FOR SALE by The House of Tabula is a 40 min investigation into the world of DIY Art, Otherness, Guerrilla Filmmaking and Outsider Art. A look into the fact that every artist is, in some form or another, an outsider.
The complete opposite of
"lofty self validating platitudes that justify strict adherence to the existing permission structures replete with nebulous definitions of "cannonized" art that have dictated HOW are visual media must be created and interpreted by the "enlightened."
You did not watch it or learn anything. You only seek conflict.
To quote Sean Connery in the film The Untouchables. " Here endeth the lesson".
The complete opposite of
"lofty self validating platitudes that justify strict adherence to the existing permission structures replete with nebulous definitions of "cannonized" art that have dictated HOW are visual media must be created and interpreted by the "enlightened."
You did not watch it or learn anything. You only seek conflict.
To quote Sean Connery in the film The Untouchables. " Here endeth the lesson".
Quote:"You did not watch it or learn anything…..
A look into the fact that every artist is, in some form or another, an outsider."
Tell me, Why are quoting this dross when you yourself don't actually practice or believe in it.
First I reject the appellation ""Artist" and never refer to myself as one,
Animator, Graphic Designer, Graphic Novelist,
"Visual Media Producer" ...hence "VMP comics"
Again what is more "radical" or "outsider" than dumping decades of time invested in learning acrylic painting on canvas fabricating my own frames for some of the paintings, with my own Mitre Box saw.
Teaching myself Desktop publishing software well enough ( after work hours) to become employed in the print design industry
( Adobe photoshop ,Illustrator,,Page maker,Quark Xpress)
later self learning 3D animation in C4D,Poser,Iclone,Blender,
And here (RETIRED) in my early 60's ,switching entirely to a controversial new paradigm shifting technology such as generative AI for animation, graphic novels even some lite graphic design with Grok agent mode.
I am the ultimate "outsider" while people like you angrily insist that creators like JSFILMZ and I conform to your permission structures on what software production methods we are allowed to use to tell our escapist stories
Such utter Hypocrisy.
A look into the fact that every artist is, in some form or another, an outsider."
Tell me, Why are quoting this dross when you yourself don't actually practice or believe in it.
First I reject the appellation ""Artist" and never refer to myself as one,
Animator, Graphic Designer, Graphic Novelist,
"Visual Media Producer" ...hence "VMP comics"
Again what is more "radical" or "outsider" than dumping decades of time invested in learning acrylic painting on canvas fabricating my own frames for some of the paintings, with my own Mitre Box saw.
Teaching myself Desktop publishing software well enough ( after work hours) to become employed in the print design industry
( Adobe photoshop ,Illustrator,,Page maker,Quark Xpress)
later self learning 3D animation in C4D,Poser,Iclone,Blender,
And here (RETIRED) in my early 60's ,switching entirely to a controversial new paradigm shifting technology such as generative AI for animation, graphic novels even some lite graphic design with Grok agent mode.
I am the ultimate "outsider" while people like you angrily insist that creators like JSFILMZ and I conform to your permission structures on what software production methods we are allowed to use to tell our escapist stories
Such utter Hypocrisy.
REPLY
! REPORT
bonj
Karma: 14,817
Fri, May 15It was about films like the indie skate video's didn't you make a poser animation for for a low budget film aired at some film festivals? It was about pushing against the grain and making up your own rules.
This reminds me of feeding a baby, Go on have a tast , you may like it. It's good stuff.
This reminds me of feeding a baby, Go on have a tast , you may like it. It's good stuff.
bonj
Karma: 14,817
Fri, May 15"Tell me, Why are quoting this dross when you yourself don't actually practice or believe in it."
Quite the opposite, i do believe in the evolution of cinema over a long history of development and technical advances. You have to understand the medium and learn the rules to be able to bend and even break them. That's why it's called a craft and an art form.
Quite the opposite, i do believe in the evolution of cinema over a long history of development and technical advances. You have to understand the medium and learn the rules to be able to bend and even break them. That's why it's called a craft and an art form.
Quote :It was about films like the indie skate video's didn't you make a poser animation for for a low budget film aired at some film festivals?
Yes and at that time many in the Lightwave and C4D community warned that I should NOT use poser for
animation commission work because
"Poser is for losers & dweebs who only make bad naked vickie art"
But I used what I knew best at the time for making animations and My client(Blue LLamma Studios) actually loved the primitive "indie game vibe" look of Poser 6 animations rendered in C4D R11 and paid gladly and gave me the IMBD film
credit listing without my asking.
Quote:It was about pushing against the grain and making up your own rules.
Right, like dumping Blender,Iclone/Character creator,a recently purchased Cascaduer pro license, Reallusion Cartoon animator(2D) and MOHO pro 14(2D)
and using Grok and mage.space, and getting Superior results in both 2D/3D animation with cloth simulation, fluids effects smoke & pyro and producing an 80 page full color graphic novel ( already at the printer) in 22 days instead of 8 months to a year.
(Old dead methods violate my new rules)
REPLY
! REPORT