Rhino Mcp/ claude first try

Rhino AI
1

Have been poking at the Rhino MCP server hooked into claude.

To be very Clear…I am a Muppet. I am not a programmer, I have an art degree, and am a certified glass licker. I do not write code, and as such tend to shy away from anything that requires a shell window. I can draw a bit and model a bit and somehow I have managed to support myself and my family for 30 years selling doodles and making silly little models of stuff.

But…this technology is very quickly becoming the new normal and as such I am invested in figuring out how to use it as an art supply to make the art I want to make, and perhaps sell a few more doodles and silly little models.

This project blew me away in number of ways and my goal for posting it is to open the possibility of how this tech can be used for folks who are actively resisting it. (like me)

@Tommi_P_Laiho (who I first met back in my Tsplines days is a lovely human and talented designer who won an early modeling contest we sponsored.) recently sent me an unfinished Ferrari p4 model for a crit. (it’s a lovely start BTW)

I decided to use it a basis for an exercise to see if:

  1. I could even get the Rhino MCP set up and running with Claude. (this was not guaranteed for me)
  2. Claude could fill in the missing bits of the model and the scene I imagined and described.
  3. could I make revisions and iterations to get what I had in my head out onto paper.

the answer to all of this was “holy sh!t”

I fed Claude/ Rhino MCP this model in the rhino window.

image
image1486×773 170 KB

and asked it to create a lemans style promotional poster, but it needed to finish the model, create a scene that included period correct details and scenery and generate the text appropriate for the request. Then frame it all in a poster format.

35 min of watching Claude run rhino to model the parts that were missing, I got this-

image
image1035×526 190 KB

Everything is nurbs and the poster is in a locked viewport set to the aspect ration of a typical poster.

It built wheels, the missing body parts, the glass and headlights, and entire interior, roll bar, engine, radiators, and all the scenery and environment details.

I had it revise the colors a bit and clean up the text a bit.

The result while impressive, had lots to nitpick and I ran out of Claude credits to keep going so I did what any good artist does, went looking for free stuff elsewhere…

I took a screen grab of the poster and fed it into nano banana and with a very light prompt to the effect of “make this better and give me 4 versions”

I got this-

image
image463×594 370 KB

Granted two of them are throwaways as they do not fit the poster style I was going for, but two of them are quite good as a starting point for further iteration and refinement.

I particularly like the lower left as a launch pad for further exploration.

I ran a consultancy for 20+ years before coming here full time and this amount of work would have easily been a week of studio time. The fact that I was able to pull this off in under 3 hours, and get legitimate assets to show a client as a 1st pass proof of concept at this speed is just mind blowing.

Is there a ton to nit pick? Sure is. Are these final pieces?

F#@k no…

But, the productivity amplification aspects of these tools are pretty intoxicating. As a solo studio operator I now have at my disposal a modeling department, a graphics department, a photo studio, a music studio, video editing, a programming department and a marketing/ promotional department.

In effect, all the previous barriers to creating art my clients want are gone. My new challenge is using simply these tools creatively to create the art I want which ideally will delight my clients (and ultimately get paid for).

To me, that has always been the name of this game no matter what media you work in.

truly wild times we are living in.

thoughts?

2

Well, if you have no ideas and no skill, and are okay delegating your soul to compute, and are okay with wasting excruciating amounts of energy in data centres, and are okay with outputting things based on vast amounts of stolen IP, I guess it’s okay ; )

3

tell us how you really feel.. :wink:

That’s one (well respected by me) opinion for sure and there are valid points for sure. My opinion is a bit more nuanced. These tools are not going away, they are already becoming the new normal and hating on them does not serve us well as designers who want to keep working.

I see it as a new art supply, nothing more. For me it’s a better use of time to get good at this new media before I’m completely obsolete.

4

From another perspective like hobbyist

I’m still thinking it’s another tool but nothing special in other areas

For 3d cad and renders my thoughts;

"Whats the point?

This is boring…

I will defeat it!

Maybe can be used for inspiration?"

And then it just filters out of my mind because it becomes irrelevant.

All i can think of AI in everything in general is how it will only make us dumber because of lack of brain usage.

It’s like common sense, it’s something that builds through experiences and not something we are born with by default.

5

Well, you asked for thoughts, and you got mine : )

6

Delegating your soul to compute does just that. More and more harrowing studies show the devastating effects of AI usage and I can already see them in young trainees or students. On the other hand, those who can think for themselves increasingly have an advantage : )

I like AI when based on my own or the client’s material. There are already good use cases for that, for example brand alignment analysis.

7

The fact I am more interested in the rendered viewport image sort of told me everything I felt about these AI images.

Woolly Rhino was a much better use of AI. To make useful tools and share that to progress Rhino.

To be honest, your final request “make this better and give me 4 versions” to the AI sort of shows that you didn’t seem to care very much about the output either.

This is how you end up with modern automotive design. “Hey AI, give me a generic SUV which is entirely indistinguishable from any other”.

Ultimately, this is the future, and I am wrong; but I am happy to be stuck for the moment in the arts and craft movement of rendering and artistic work.

8

and I appreciate and respect them! :wink:

9

I guess I should clarify.. I see this as a very efficient way to create the “first pass” of a design project. The one where you are still searching for a common language with your clients. I often referred to this phase as a “hate model” which was a model I knew they’d likely hate, but they are not yet communicating effectively what they do want, so I’d do a pass that was deliberately vague or had obvious flaws with the intent of drawing out their direction.

A lot of clients are great at telling you what they do not want, but not so great at verbalizing what they do want.

In this particular case, this early first step process was dramatically accelerated and I got to something I could show to a client very quickly.

Like I said… is it “done” noooo.

But it’s a great start to get the details I’d need from a client to start closing in on what would be “done”

Did I care about this particular output? not really- but it’s still an important step in the overall process. I’d fully expect the client to make a ton of notes and we’d iterate several cycles to get to “done”

with my clients it was this first round of development that took the longest amount of time when it came to getting a client to sign on to a direction, this could be a jet pack to get thru that step.

my goal with all of this is to offer a different way to look at this stuff and have a dialog about how it can be used to benefit us all instead of replacing us all.

keep the thoughts coming, I appreciate the conversation, even if we disagree.

10

we had a similar “holy moly” moment in our office a week ago using Veras / Enscape.

one colleague, who has no formal videography or animation experience, made a 90-second sizzler reel in about 3 hours. with moving people, moving bikes, moving cars, wind-blown vegetation (and people’s hair/clothing!), weather (clouds moving, sky changing colors), daylight, electric lights. the reel had several interior, exterior, inside-looking out, outside-looking in shots both in daytime and nighttime with transitions from late afternoon to night.

the project is entering construction, so the sketchup model we used is pretty complete with entourage, vegetation, site elements.

i just know this type of reel would have taken weeks, maybe a month or more to produce working with outside partners. and outsourcing still an option, it’s not like we’re claiming our reel is as good as a team of artists making art with our project. in fact our sizzler reel is just the sizzle of the steak that a team of artists would create (grill?) if the client wants a T-bone to chow and is willing to pay for it.

so it’s not a replacement of craft, but really a way for the client to taste something they may not be familiar with before they buy the real deal.

11

Moronic take.

No ideas and no skill? What has that to do with anything?! @theoutside Is highly skilled and he could have easily done this but it would have taken a lot more time! That is the whole point. Also, as you can see from his experiment, he create something a team would need 1 week, skills from graphic design, to 3d modelling to editing and stuff, no person has every skill in the world. What the f do you mean if you have no skill? LOL

Also, nothing wrong with having no skill. You sound like an old painter talking down on digital artists. Like sure, of course if you have no skill to calculate perspective, lighting, shadows, reflections and colors, then I guess 3D is okay and render engines are okay, oh and also you waste so much energy when I just use my paintbrush and my skill. :rofl:

3D opened the door for millions of people with nor formal training in arts to become artists. Nothing wrong with no knowing how light affects objects. That is what the render engine is for.

Why do we even use computers at all? Why do we use internet at all?!,Let’s go back to letters, candles and manual work. No electricity, no energy wasted!! :rofl:

At least I wouldn’t have to read moronic comments made on a PC which wasted a lot of energy.

Also “delegating your SOUL to compute”?! :skull: :rofl: :rofl:

What is bro talking about.

Bahah, I am sure they are, very proven and concluding studies.

Same way internet made people stupid because you could google everything right?

12

What future? Choose any color as long as its black? That was 100 years ago, and you are saying its the future? Scared of the generic?

13

remember we all friends here…

all opinions are welcome, even opposing ones.

lets hear from everyone, with no name calling please.

Your take is valid as well thanks for taking the time to chime in.

14

for the parts of the model that Claude built, do they seamlessly fit into the geometry you provide? in other words, could you join the windshield with the body into a single polysurface?

15

Whether someone chooses to leverage AI—or where they decide to apply it—is entirely up to each individual or company. However, I believe that understanding it is both crucial and essential.
I truly hope everyone keeps experimenting with it.

16

not as drawn, but it would be an easy repair to make to the geometry that it created.

I have not tried it, but I would assume if you called out those surfaces it could refine them to your spec.

17

You must be fun at parties.

18

Can you back that up with data?
If I look here it seems drinking tea all day is much more of a waste.

19

Informative article to start with, though I guess it should also include image and video queries in the study.

20

Absolutely, here’s a good start. No wonder that ever more citizens in the U.S. and Europe are protesting against the U.S. techno-feudalists plastering the landscape with energy consumption intensive data-centres. Also, the “data centre build-out = high quality well paid job growth” mythology has been frequently debunked.