Runchat AI rendering

Gallery
,
41

I think something that might be of note in the whole AI debate is how to look at this from a forecasting point of view. We are very “in the now” creatures, so we tend to react to what’s in front of us but don’t do well at really seeing things from a future perspective (naturally since we’re not fortune tellers).

So the closest thing we have to preparing ourselves are tools like “hype cycles” and forecasting charts, which is part of what my job is about. So how that affects us is in the sense of what is coming in the next 2-5 years (or likely to).

  • Is it likely design, engineering and AI are going to be embedded? Yes, it’s already happening. How remains to be seen.
  • Is it likely AI and storytelling for pitching and selling our designs is going to be more common? Yes, it’s already happening at design school levels and bigger design pitch decks, marketing productions etc.
  • Is it something we need to adapt to and start using now? It’s debatable in what way that’s going to play out because we’re still at a stage where AI can’t fly solo on alot of things or are not ready for public acceptance yet (plus regulatory things that are starting to happen in reaction to some “whoa whoa hold up, we need to slow down” reactions from key players.

What I think is certain is that once the hype cycle dies down, the dust settles from novel/pointless AI companies dying out, we’re going to see these new hybrid workflows become more accepted and even nessesary. We’re in a quiet recession, money is tight and now is really the type transformative moment companies are using as an excuse to slip in AI into things we do see and many things we don’t see.

Conclusion: Whether we like it or not, I think we’re coming up to a moment where many traditional workflows are going to be forced into some type of AI workflow or assistance as more and more people expect it. The question will be is how does a company think about aligning with that (or not). and how much of a gamble will that be.

For example, Procreate has taken a strong stance against AI and some argue that could be a very detremental thing down the line to company if customers are expecting AI assisted features like something like photoshop already has.

image
image1223×1125 63.5 KB

42

Gwylim shared some credits for playing with. Would love to hear what you learnt from the Runchat team. Im running Rhino on Macbook Air, so cycles and bella throttles the performance. Remote rendering via an AI plugin seems to be the best solution for me.

43

so far my testing has been only a few sessions, but I am impressed. I also feel like this and other AI based rendering platforms like Vizcom is likely the future of rendering due to it’s speed, ease of use and results. I think there will always be a place for fiddler rendering engines like cycles and keyshot, but for a daily driver, this ai assisted stuff is IMO already really hard to beat. Especially for early concept work and iteration.

44

The dismissiveness is remarkable.

In this age of amazing modeling and render engines. Generation of geometry with Grasshopper and Rhino, what’s available in real time interactive visualization with Unreal Engine. The tools are beyond what we’ve wanted our whole design lives. Yes they require effort and expertise.

Yet here we are, playing with the automated glazing machines. They tell us our ideas are great, they render our rough concepts into glossy sexy pixels.

Its a question of control and effort vs phoning it in. Prompting it in?

The bean counter “monkeys” at the monkey show, they know about Runchat and Vizcom also. They have hundreds of thousands, or millions of dollars at stake to produce a product. They are not going to be “sold” on a bitchin’ render they know/assume took max twenty minutes to iterate and a cool “context video” made of six second clips. They will want to know about the provenance of a concept and it’s execution. They will suspect every design produced in 2026 and beyond will have been done by prompt. They live in the same era we do.

The glazing renders are cool, the magical summoning is cool, seductive even to the designer. No need to commit to working something out, just rough out something and the glazing engines will make it look slick af! And look, a cool video of an attractive person using it!

Skill/output confusion. The ability to generate superficially impressive visuals becomes conflated with the self perception of design ability itself. The tool’s rendering capability and hole filling gets self-attributed to the operator’s design judgment.

What’s at stake isn’t whether AI rendering tools exist or improve, that’s inevitable. What’s at stake is whether the profession and producers maintain any meaningful distinction between surface execution and substantive design work.

45

Yes this is what I call my “kick ball, cash check” approach. :rofl:

46

This is a concern I have always had and one of many reasons I wont use it. I dont want the validity of my work to be compromised due to the assumptions of the uninformed. If someone asks me if I use AI to create an image I want to be able to say no. But I also know going forward that it will start to be just assumed to be AI. How do I know this? Experience. As Ive stated earlier I was in an industry that has been replaced by CGI. During the early days of that transition when CGI was being marketed heavily as the way to go I had much of my work that were physical builds shot on a stage just be credited as digital. These days when I point out work that was done practically people dont believe it until I show photos of me standing next to something on a stage. Been here , done that.

47

And then, there are those who will ask you:

“Did you create this using AI?”

And you proudly answer “No!”

And they think, “Oh great, a Neanderthal from the last decade. All we want is results, and we do not care how we get there.”

(that last reminds me of a line in 3 Days of the Condor, talking about oil. “When they start getting cold, they won’t care how we got the oil.” Or something to that effect.

48

Im reminded of a line from Star wars…..

49

I am enjoying this conversation, and offer this thought.

if we removed the controversial word “Ai” from the conversation, and just looked at the process:
-a person input some stuff into a computer and got an image as a result-

what difference does it make what was software used to make it?

Does a hand drawn image have more artistic validity than a very nicely created Keyshot rendering?

Along those logic lines, does a hand drawn image on paper have more artistic validity than one hand drawn drawn digitally in procreate or sketchbook pro? (gasp! there is undo!!, panic, there are vector lines!, swoon, there are color manipulation tools! THE ROBOTS!!! THE ROBOTS ARE COMING FOR US…. ALLLL IS LOSSST!!!)

no… we don’t say this about procreate, photoshop, illustrator or any other digital art supply. No one even blinked when that stuff came along. Wacom tablets? they are “artificial paper”… no one cared for a single second.

We can argue that AI is nothing more than a digital processor…it’s not sentient, it’s simply a very very large, very cleverly put together search engine that just happens to be wired up to a very clever image generator. (among other things)

How is this any different than the very clever digital processors we call unreal engine or vray or keyshot? They take input from a human, and deliver results that are very clearly “artificially generated”…no one cares. In fact they are celebrated in many circles.

My experience using this is identical to the experience I have when running any render engine or any digital art tool for that matter… It just has a different UI and faster results. I’m still making the art I want, getting the results I envisioned in my mind.

Your mileage may vary, but I’d caution all of us against falling for the “sky is falling” mentality surrounding “Ai” and examine the potentials to accelerate and improve your own work.

The reality, Design is a biz that is not kind to those who don’t keep up with the times.

Thanks for all the thoughts and contributions to this discussion…I’m learning from you all everyday.

50

So, if the continous cycle applies; following this logic, then there is absolutely no need for McNeel, nor indeed Rhino to exist. Bluntly, there is no need for your tutorials, nor this forum.

At this point, we may as well skip the design process and publish directly using whatever globally averaged results are provided by the RTX monster server at the click of a button. I think I could gather an aggregated query such that your initial design input is also utterly pointless.

I say this in the same way I see the evolution of car design. AI is much like the advent of globalisation of cars; it removes all sense of culture, the real need to communicate, and understand the tool properly.

The near instant gratification will inevitably make the artistic world more numb. Just like CGI is now far too much in films, to the expense of the storyline; AI will just feed us so many results with hardly any effort, all meaning will become lost.

Thousands of years of technology, improvement in medical care, sciences, agriculture; and what that has meant in the last 5 years is that we have invented a technology that causes us to not even be bothered to defend or put effort into one of few things that makes us genuinely distinct as a species.

The difference is very real for me. A clay model, a chisel, a Wacom, a rendering engine… all have had people spending hundreds of careful, considered hours to get excellent results that people want to learn from, that people want to talk about, and people want to care about. Learning AI prompts is just optimising how best to walk around the exam room looking at everyone elses answers, and taking an average.

I really miss the days of getting home from school or work, and wondering what amazing thing that guy has done on his space ship model today. Can I learn more?

51

I was going to respond but I really cant put it any better than that.

52

@David53 Exactly, thanks for the summation.
Seems obvious with the slightest bit of introspection.

It’s deliriously precious to mock experience and the craft of doing something right.

“Did you use a designer?”
“No, too expensive, we used AI. Jerry from accounting did the work in-house on Vizcom, cost us next to nothing.”

“Did you use 3D software?”
“Nah, we got a guy with a Rodin sub. Cheaper to fix it in post than pay someone who knows what they’re doing.”

“Do our customers want AI designed products?”
“All we want is results, and we do not care how we get there.”

Let’s see how this works out, we will know soon enough when the first generation of AI products hit the shelves.

To care or not to care, that is what it boils down to. Reputations are built on both approaches.

53

I completely agree with David!!

54

I consider the AI rendering just a tool as every other tool.

And I appreciate that acceptable renderings can be made without having to know every nitty-gritty detail about the properties of light, material etc.
When I remember Brazil, which could produce excellent results, the practical work was more than cumbersome, it felt like rocket technology – at least for me.
Also there was Flamingo NXT, much easier, the possible result were also good, and I didn’t feel stupid when using it.

AI rendering is even simpler to work with, that’s good.

On the other hand I miss the human touch in AI results.
They feel too artificial – no wonder.
Same with AI “music”, there’s no soul in it.

55

”Where have I heard this conversation before… oh, yeah”

image
image1825×1373 348 KB

56

I don’t think that existing viz pipelines and new tools necessarily need to be mutually exclusive. Instead I think there is an opportunity to get creative with these tools and remove grunt work and friction that gets in the way of play and exploration. Here’s an example of generating tileable textures (some cliched timber planks, sorry) from a reference and then generating the code to download the images, create Rhino materials and assign them to selected objects.

I don’t think that the skill here is prompting per se. And I don’t think that the goal is to generate average outputs. It’s about finding ways to design or communicate that feel more intuitive and natural. I don’t want to get into photoshop and make tileable textures. And I definitely don’t want to search the internet for textures - I just want to provide an example of a finished product I had in mind and then work backwards from that.

It isn’t too hard to imagine how these hybrid approaches might extend to delighting reference images, generating multiple versions of textures for less repetition, etc etc.

57

Your thoughts remind me very much of the new TV series Pluribus that began a few weeks ago. The whole story is unkown at this time. So far- A genetic sequence is broadcast and intercepted from outerspace. Scientists create the sequence and it quickly infects the whole earth turning all but 13 people into an interconnected and coordinated network of intelligence. I becomes “us”. All intelligence from all people is instantly accessible by all infected people. The genetic infection can’t destroy or hurt anything. It only wants to assimilate and propagate itself. I see it as a kind of AI. Most of the uninfected love the new orderly world where their every wish is provided instantaneously by the hive mind. Only one (or two?) of the 13 sees this as the end of humanity and she is on a quest to reverse the infection before the hive mind discovers a way to infect her and the other 12 into their “joyful” productive life.

58

good point, at the moment i feel better entrusting ai to make images/text (symbology for use in ideation) than i do entrusting it to make accurate g-code or even reasonable digital models (symbology for use as instruction).

applications like what’s in the smarthopper thread seem neat because it’s integrated into an environment in which you already work, and there’s more of a back-and-forth between your work and its work. it’s not replacing the experience that you’ve gained over years (unless you literally take it at face value and don’t check/edit its work in which case you have bigger work ethic issues), but helps to resolve well-defined known unknowns (“i want X to do Y via Z”).

59

Imagine we had a robots with built in AI and you didn’t need to be a carpenter anymore…you could just tell the robot what you want and a few hours later you’ve got a wooden cabinet rich in details.

Then you would say: “If we removed the controversial word ‘AI robot’ and just looked at the process - a person uses a machine to make stuff - what difference does it make what machine you use?

Just because two things are machines, or two things are software, doesn’t mean they’re equivalent.

You don’t think we’ve lost something when more and more we’re better at prompting AI and worse at actually using creative tools ourselves?

60

one point of the Pluribus story you missed…

the main character is an utterly miserable human and is actually fighting for her “right” to stay miserable.

The debate around that specific topic is one of the wonderful plot points in the story.