Introducing Blender Buddy, your AI expert Blender mentor

General Forums Off-topic Chat
1

Meet Blender Buddy, your around-the-clock mentor for mastering Blender! Designed as an indispensable AI assistant for 3D artists, Blender Buddy helps how you interact with Blender. Trained meticulously on the in-depth Blender manual, it possesses an encyclopedic knowledge of every function and capability within the software. But its expertise doesn’t stop there – Blender Buddy also integrates wisdom gleaned from years of expert questions and answers, capturing the nuanced challenges and solutions experienced by seasoned 3D professionals.

Blender Buddy stands out as a 24/7 mentor, always ready to offer guidance, tips, and insights, whether you’re tackling your first project or pushing the boundaries of 3D art. Need help with intricate modeling challenges or optimizing your renders? Blender Buddy provides instant, context-aware advice, any time of the day or night. It’s like having a seasoned expert by your side, constantly available to help you hone your skills and elevate your creations. Embrace the future of 3D artistry with Blender Buddy, your ever-present guide and mentor in the exciting journey through the world of Blender!

This is using ChatGPT’s (You have to have an OPEN AI paid account).

This is something I wish I had many years ago when I started.

Ask any question about Blender

*** You can upload a screen shot with some circled edits on it and it will analyze the image and what potential problems you may have***

*** It will also write Python Code for you to help with design***

Hope its helpful to some.

https://chat.openai.com/g/g-fXcjWY0nd-blender-buddy

I AM NOT SELLING ANYTHING!

I will make this better as time goes on. Still many things to improve.

Once I have time and money I will recreate this with completely open source models

*** KEEP IN MIND THIS IS AN LLM, THEY ARE PRONE TO HALLUCINATE*** ANWERS MAY BE WRONG. I will work to minimize this.

2 Likes
2

blender-buddy-taking-some-time-off

3

It looks like OPEN AI hasn’t rolled out GPT’s to you yet. They are slowly opening it up to everyone in the coming days. So it should be available soon to you.

4

Hmm :thinking: … don’t we do have this already ??? It’s called BlenderArtists Community… ask whatever you want… :stuck_out_tongue_winking_eye:

…even if it is already in the docu… or mentioned last week… :wink:
5 Likes
5

Some interesting use cases.

*** KEEP IN MIND THIS IS AN LLM, THEY ARE PRONE TO HALLUCINATE*** ANWERS MAY BE WRONG. I will work to minimize this.

Screenshot 2023-11-10 at 11.45.04 AM
Screenshot 2023-11-10 at 11.45.04 AM1746×2074 660 KB

Screenshot 2023-11-10 at 11.48.38 AM
Screenshot 2023-11-10 at 11.48.38 AM1672×1834 524 KB

Screenshot 2023-11-10 at 11.52.18 AM
Screenshot 2023-11-10 at 11.52.18 AM1724×2038 782 KB

*** KEEP IN MIND THIS IS AN LLM, THEY ARE PRONE TO HALLUCINATE*** ANWERS MAY BE WRONG. I will work to minimize this.

1 Like
6

These are interesting examples, they really show how dangerous it is to rely on something like this. Examples of things your buddy got, confidently, wrong:

  • In your first screenshot, is the origin a yellow or an orange circle? Blender Buddy says both.
  • On your second screenshot, you definitely don’t have the outliner circled. You don’t have elements circled at all, you have one element circled.
  • “.[individual origins]… touching faces are treated as a single element”; no, they aren’t, each individual face is treated as a single element
1 Like
7

Yea I noticed that. But its a great starting point and pretty impressive. If a beginner could just post a screen shot and ask questions about it. That would be very helpful . Yes initial answers are a little wonky but its getting better by the day. As I add more example and things not to do this will be an extremely helpful tool. I also can find other errors you didn’t point out, but thats not the point. It is trying with very little guidance. Its a million times better than waiting weeks or months for an answer on stack exchange and going through everyones BS negativity on BA.

PS I cant believe you said “DANGEROUS” HAHA you sit behind a computer pushing vertices. I can assure you nothing you do could be considered “dangerous” to anybody.

8

…why would someone ask “please tell me what is on page 67”… instead of just reading … page 67…

9

What if I didnt understand what page 67 meant? What if I want to make sure we are on the same page!?!

10

For some people, this is a hobby. For some, it’s a job, and their livelihood depends on it. Dangerous doesn’t just mean physically

1 Like
11

I have been using Blender professionally for over 10 years. I am busier than I have ever been. Not sure of your point. If AI puts you out of business I am afraid you have other problems to worry about.

12

Things can be “dangerous” in terms of social responsibility.
Misinformation is “dangerous” - it interferes with person’s education, forms misunderstanding, wastes precious time. Especially when the target audience are impressionable beginners.
Systems like that are under some obligation to be either VERY good and precise… or not deployed at all (imo).

13

This is absolutely ridiculous. Yes of course if this is for Medical or Financial advice. This is for people creating art on a computer.

Also I can assure you that there is A LOT of misinformation on BA and SE. Should these also be shut down then?

14

I think he meant that sometimes being confidently served as an answer something that is false can cause harm, particularly if you depend on it for a job.

LLMs are cool, but the fact that I can never trust an answer not to be completely made up is in itself a waste of time for me. Having to distinguish what might be correct from the crap is an extra layer of effort, at least for technical things.

4 Likes
15

I may have unintentionally opened up a can of worms here, I didn’t mean to derail this discussion. I just wanted to point out that there are mistakes here not pointed out, and I should have left it at that. That said, I do believe it is your responsibility to acknowledge mistakes in your product when sharing a product, and while it would have been better to do that before someone else did, I’m glad you are willing to acknowledge them :slight_smile:

1 Like
16

It’s not that it will put people out of business, it’s just misleading and confusing sometimes. Being told to look for a yellow circle when the circle is orange (and god help you if you’ve customized your UI) is a waste of time.

I see this like an as seen on TV garlic peeling gadget. It does a great job on the cloves that are already easy to peel, and the ones that are harder to peel, it does nothing.

For easily googlable questions, it offers nothing that googling the same question wouldn’t. For anything more complicated, it’s not going to help at all.

I use chatGPT for python development, and it’s great for simple tasks, but if you ask too much of it, all you will get is error after error, with no clear way of how to solve the core issue. When I actually take the time to write code, then I know how to fix the code that I wrote.

That’s why it is dangerous, it lulls noobs into a false sense of security that blender buddy can solve all of their problems, when it definitively cannot. Once they do run into a task more complicated than adding a suzanne model, they have developed zero problem solving skills.

But if you are just looking to exploit new users with the lie of an easy fix to a complex problem, then scam on, I guess.

Also, LOLOL at the “in-depth” blender manual. You can’t swing a dead cat on these forums without hitting a post complaining about the incomplete manual. I get that you “meticulously” trained the model on the manual, but garbage in, garbage out. All you’ve done is make a high-pressure garbage extruder.

5 Likes
17

It’s true, the manual is mostly worthless. I’ve learned that you’ll find more accurate information anywhere else at this point. Of course, a major portion of that is that the Foundation doesn’t have someone working on just the manual, and there’s a massive backlog of typos, errors, and outdated information that gets worse with every commit. They need to refactor the whole thing, but that would be an insane undertaking.

Anyway, as you were :slight_smile:

2 Likes
18

NVM the negativity on these forum has been getting worse as the years go on. I have been a die hard Blender head since 2.5. I fully understand why outsiders cant stand this place.

Its a tool that could be helpful to someone somewhere. I made it for free with hours put in to gathering data. I am NOT making money off of it. Just trying to help out.

Use it or dont use it.

1 Like
19

Absolutely, and I’ve likely been the source of it more than once. But thankfully, there are others reading it, and cross-checking it in their head, and correcting it when necessary. So the bad info tends to die properly.

Chatbots don’t do that. They present the information as correct and absolute, and if they don’t actually know the correct answer - they’ll fake it to the degree that the reader doesn’t even know it’s wrong.

No one is saying you’re a bad, stupid, evil person or anything negative about YOU whatsoever.

Dear god, when did it become a freaking social taboo to not accept anything and everything as a great solution? When did saying “I think there’s a problem with this” become always taken as a toxic negative attack?

4 Likes
20

Not sure why you are speaking for others (some alluded it could be a scam, smh). I wasn’t addressing you at all.

My point is people come here excited to share ideas, products, stuff… and this site just takes the air out of everything. Other Art / 3D forums are NOT like this. (*** Here comes all the screen shots from other forums ***)

OK I am done talking about AI and its Dangers, thats another thread I have been avoiding.

I will show cool (possibly wrong) examples in the future of something neat I was able to make it do.