I know Blender AI is a divisive topic here, but this is more of a Blender Foundation news thing than an AI hype thing…
Anthropic (the company behind Claude) just announced “Claude for Creative Work”, a broad partnership across creative software. They’re releasing connectors for Autodesk Fusion, Adobe Creative Cloud, Ableton, SketchUp, Splice, Resolume… and Blender.
The Blender side is what caught my attention. Two things happened:
1. The Blender developers built an official Blender MCP connector for Claude. Not a third-party project. The Blender team built it themselves. It connects Claude AI to the Blender Python API for things like scene analysis, batch operations, and custom tool building. It’s an open MCP server, so it’s not locked to Claude, other LLMs can use it too. If you’ve seen 3D-Agent or the open-source blender-mcp, this is the Blender Foundation’s own take on the same idea.
2. Anthropic joined the Blender Development Fund as a patron. Real money going into Blender development, specifically toward the Python API and AI.
Now, whether you like the Blender AI direction or not… the Blender Foundation clearly does. They put Blender MCP and AI integration on the roadmap in 2025, and now they’re building official connectors and accepting partnerships with companies like Anthropic. Blender is listed in the same announcement alongside Autodesk and Adobe as a creative platform. That’s just where things are going at the Foundation level.
Practically this doesn’t change much day to day. It’s an integration with Claude’s products. But the funding and the fact that Blender keeps showing up in these partnerships is worth noting.
Thoughts? Do you think BF will regret this move since most users are against AI?
Between that, Clement cancelling the NPR build, and the utter failure of the Eevee team to fix serious known regressions they introduced with Eevee Next 2 years ago - I’ve cancelled my dev fund donor status.
It’s not very much money at all, and I’m sure they won’t miss it (nor really care) … but it’s about the principle of the matter, more than anything.
They’re taking Anthropic’s money, and Andrew Price’s money, and MeshyAI’s money… so, yeah - I guess they’re covered in the funding department.
And I’ll be honest… even though I’m positive about AI tooling in Blender, Anthropic specifically is a weird partner for the BF. Blender’s whole identity is open source, community-driven, free for everyone. Anthropic is a closed-source AI company, to say the least… Those philosophies don’t exactly line up. I would’ve expected this kind of move to come alongside open-source LLMs, not a proprietary one. The fact that the connector is built on MCP (open protocol) helps a bit, but still.
I saw the post on the Fediverse stating the Blender project is now getting sponsored by Anthropic, and would like to voice my views on the matter. I wanted to voice them on the devtalk forum, but when asked where to voice them, I got a non-answer.
The project does state that this sponsorship will not affect the stance of the Blender project, and even if that is true, the Blender project already seems not to have a negative stance on GenAI, which to me is horrifying.
This technology is fundamentally unethical, it’s built on a lot of violence and has the express purpose of replicating human creativity. It strips humanity from everything it touches and devalues humanity itself. The project, in my opinion, should have a full ban on GenAI at the very minimum. For code, promotional material, assets, artwork, 3d models and everything else.
It’s also important to be in solidarity with everyone affected, so we can have a unified front against all GenAI. With artists, writers, programmers, and everyone else.
There are many more reasons why this technology is fundamentally unethical, including resource use, effects on communities where datacenters are built, mass misinformation, and many others.
I feel like the minimal morally defensible stance is a full ban, and it would be preferable to do more than that, like actively spread the message that this technology is not acceptable anywhere.
A policy such as “GenAI outputs are not allowed on Blender or any related projects. We will not knowlingly accept any GenAI outputs and will revert whatever we can when a GenAI output is found in the project.” would be great. The minutia could be discussed later on.
Hope I’m not posting this in the wrong place, but do want to add my voice to any discussion on it.
I’m not against AI itself, or generative AI, the problem is when copyright protections are violated, it’s not the AI’s fault, it’s the people that break these protections that are at fault. For artists, you already cannot take somebody’s artwork and use it without permission unless that art is creative commons. If somebody wants to use your art to train their AI, you should be paid for it, the same way you are owed money if they use it for anything else.
Remember the “you wouldn’t steal a car” anti-piracy ads? And then these companies downloaded the entire internet.
I tend not to use the term “AI” as it got so overloaded that it barely has any meaning anymore. I personally don’t care about someone using an image recognition model or a Bayesian spam filter or whatever. What I am specifically against is GenAI, or, generative models.
The copyright angle is of course one of the angles you could use to argue against it. I tend not to argue from that perspective as I think copyright is used as a proxy for a more fundamental issue, and that is hoovering up people’s work without consent and then soullessly replicating it.
This doesn’t mean I don’t agree. I feel like, if you are training an LLM, diffusion model or such on code, writing, posts, profiles, artwork, 3d models, etc of anyone, you need their explicit consent. This makes every generative model that currently exists very unethical. Including local ones.
Though, I tend to go further, because the whole purpose is to regurgitate and replicate human creativity, and this is, to me, offensive.
That’s fair enough, I get there are philosophical sides to this issue as well. I just think there are already rules that prohibit using people’s work without permission, and they should be enforced.
The policy states that this site is for humans to share words and art created by majority-humans. It covers everything from art to add-ons to forum posts. If a human did the majority of it, it’s allowed
Well, that kind of biased comment doesn’t help to make an healthy conversation.
hobbyist, beginners, working or unnemployed you can find them in 3 categories regarding AI.
I guess in advertising AI is more used, in animation I didn’t see it yet.
And it’s not two clear camps to me, some people have a radical opinion toward or against AI, and I’m not sure any of their vision will actually work in the long run.
In the other hand, these tools are part of our life now, they have great potential yet they have their innerent flaws unlike what people promoting AI want us to believe. Discussing that is way more interesting to me than trying to make a “us vs them” debate. We all live in the same planet at the end of the day …
I’m sure it’s a mistake you should definitly report it and I’m sure it will settle.
I’m having a hard time thinking about someone ripping your code with good intention. Indeed they may try to add malicious code after an update or else.
I don’t think it’s that they don’t care, maybe it’s just the best balance they found.
Extension platform is just an attempt to centralise ressources around blender, and it’s likely to be beneficial for blender and the community. But it’s just a subset of what blender is. And the goal was also to have extension as a community responsibility rather than being supported by BF.
Obviously they are much more cautious with code review that goes into blender, because it’s the core of their jobs…
Hello and welcome here !
Personally I don’t see GenAI applied to imaging exactly the same as GenAI applied to programming.
If I post some artwork of mine I consider that my intelectual property and I wouldn’t want that to be ripped of in any way.
If I post some code or an answer to someone online, to me it’s kinda participating to the overall knowledge base. I don’t mind if you take what I’ve wrote and use that for your own work even making money out of it. If that’s an issue for me I won’t post in the first place.
It’s also why I might react differently to AI assistants. If the claim is that we can take all the work of people to eventually replace them by AI , this is definitly wrong.
Now if we consider these tools as assistants, a shortcut to some result, or automation I think it’s a different stance.
Let me give you this example that is unrelated to AI but it’s also a bit similar to me : To make a complete character rig from scratch it takes at least 1 week.
Now we have autorigs that can do the same in a few hours, these autorig are based on the knowledge of other riggers and probably the study of hand-crafted rigs, books about rigging and so on. Yet on some cases they make riggers useless. Or as a rigger instead of having 1 week or more of work, you might now have 1 day of autorig + 1 day of small additions that the autorig can’t handle.
Does that looks like a fair situation to you ? Yet we are now crowded by assets / addons that are supposed to make us more creative and in the meantime it makes artists a bit more useless in some cases. Is this a good or a bad thing… I think it can’t be a simple answer…
Back to AI, I’ve used it in programming to save some time, and it was pretty helpful : I think I wrote 80% of my code and the 20% wrote by IA was pretty technical to a point it would have added extra weeks to figure this stuff out.
In this situation I don’t think AI is making programmers obsolete but it’s a powerful tool in some situation. And I think the way MCP is introduced by BF is heading toward that vision and to me it’s a good move.
I’ve seen a lot of BS MCP for blender that do nothing interesting nor useful and they claim that you basically don’t need to learn 3D anymore and that’s really a scam to me.
Now the fact that you can prompt an AI to try to clevery rename all your objects in the scene looks interesting. Just like being able to run a diagnose on the scene stuff like that. I doesn’t make art for you, it’s not a replacement of learning blender, it’s just a different way to interact with blender that in some cases can make sense.
I think it’s better that it’s done by BF with still a strong position about AI as an art generator rather than having different individual surfing on that hype with less ethics.
So some code of yours that you are required to license under GPL or GPL compatible license got reused in another add-on that is also licensed under GPL? So? What’s wrong with that? AI could spit the same code out and then it would be a problem if it’s used in software not licensed under GPL. That would break the license. This seems to be a rare case where it works fine. Or was it something else? Was it not only a snippet and your name got removed from it or something? But to be fair this sort of GPL violation happens all the time without AI as well. I don’t see this as the best example of AI breaking copyright laws… It’s not even clear that GPL license was definitely violated here.
When the moment comes that one-too-many developers have decided, “You know what? I’ll just stop sharing my code with the world, and keep my tools for myself” - enjoy it.
For that matter, does Blender really need to pay Nika to review extensions any longer? Claude can probably handle it, right?
Maybe they can also stop paying Pablo to make all those “work plan flowchart mockups” … Anthropic can do that too with a few prompts.
The Triage team has already been trying to add AI to their team - who knows, maybe Claude will fill that need completely, and Blender can save even more cash on employees and just let Claude handle the bug report filing.
I’m sure the Blender Studio is safe, though… since AI imagery and video isn’t really a thing…
Well… We are talking about GPL… Don’t use it if you don’t want to share.
That would be sad. But I don’t think these ideas exist amongst Blender developers. They might be doing stuff in their own way as Joseph mentioned, but they don’t seem to have plans to use AI for reviewing any time soon. At least that’s the feeling I am getting very strongly.
Hey I think you’re definitly not alone in that situation… In the meantime I think AI is going to challenge us to be more creative and original…
If you look before AI there was some kind of a generic look/style we were after :
In animation you had a generic pixar-like design that was taught in school. In videogame there was some kind of arstation-like aestetic that many tried to reach. Or you could be struggling to get that anime look to your 3D models.
At that time if you managed to get into some kind of average of these style you where basically good. This is what was expected in art comunities and in the professional world.
Artstation was already considered some kind of a self-regurgitation and so are classic pixar aestetic…
Now some folks came up with a technology that can regurgitate all that even better that what we can do…
And while this tech is really good at making an average of what everyone was doing, it came way less competent when it comes to originality.
That IMO, where we have to look if we want to stay relevant…
Back to compelling stories, interesting designs, focus on creativity rather than tech stunts… And maybe collaborate more. Like maybe if you take 5 experts in their own field and let them work together you’ll get something really unique that no AI can come up with ?
Anyway, the seed of the problems we face now with AI in the art community were already there before in some forms, but it was still manageable. Now AI is likely to force us to shift mindset and adapt to that situation and that can bring cool stuff in the meantime !
Hey, no, we’ve been over this. It does not matter what you can legally do with the GPL license. Stealing add-ons is piracy and it’s not allowed here. Your feelings about it are not relevant