Here is a big question—why would it get better? Presumably they scraped all the training data already. 5 million cars doesn’t mean you got the ability to go to the moon.
I guess I don’t care if it gets better as long as I can get better/faster/ and more profitable.
the tools are just tools, just like keyshot or photoshop (which were also much bemoaned when they came out) It’s up to me to use them in a way that makes me a better artist.
I get the hate..I do.. but… I have also just recently seen a peek at what I can do with it and I see a path to using these tools just like I use rhino, photoshop, or any other tool to make better art.
remember when photo retouching was done via airbrush?
I do… IMO this is just the new airbrush.. and I’m not going to quit a career doing stuff i love because the tools changed.
thanks for the ongoing and very interesting conversation… I appreciate you all.
Not being really interested in AI, I only recently started thinking how to integrate it into my workflow and mainly because I was asked to do so at work.
I am not a fan of subscriptions either so I was searching for something open-source that could run locally on my PC.
I’m using a software that is similar to Runchat as it uses nodes to increase the control you have over the output, it’s called ComfyUI. (https://www.comfy.org/)
For some tasks I’ve actually found it useful, for example, I design metal accessories for the fashion industry, and a workflow I downloaded into ComfyUI allows me to get a grayscale depth image from a simple photo to use for embossing, all in a matter of seconds.
For rendering and image generation
Paul Hansen - Archviz x AI Free ComfyUI SDXL x FLUX Workflow - Showcase
Other interesting uses of AI, for local chatbox I use Ollama (https://ollama.com/)
Arturo Tedeschi - Connecting Chat GPT with Grasshopper
https://www.youtube.com/watch?v=gL3B5KwZq2M
I have mixed feelings too regarding the use of AI, at least in the use cases of this post it feels more like a tool.
For my personal projects I prefer doing all the work myself simply because I enjoy doing so… modeling, setting up scenes for render etc. but it’s always interesting to me trying new tools.
My thoughts are that you are, unfortunately, probably correct. I find myself very much in the anti-AI camp, but I realise that there’s no putting the genie back in the bottle
A couple of years ago when ChatGPT first appeared I remember thinking “Huh, yeah, that’s pretty clever”. Then I watched someone make an image with DALLE. I felt a pit in my stomach and my mind turned to thoughts of “this feels like the end”
Moving forward, I have in recent days been thinking that it might be best for me to pick up more non-digital interests. Not just for engineering/career sake, but I reasoned that any art, gadget, or physical entity that actually occupies real space and is going to be more AI-proof that a digital art piece
Those of us in the engineering world will probably adapt as has been required before, and AI might become a welcome tool in the utility belt. Personally it’s the artists I am most concerned for. This year a parent company I was working under registered two new SPEs. Previously they had hired an artist to design the company logos. This year they just generated logos in a few minutes using AI
(It was obvious AI slop too, the logos just had that “look” about them 
)
The problem no one seems to mention is that ai aint free. It’s a scam to nickel and dime people and it’s succeeded. Most ai gives you a certain amount of free tokens then it charges. I’ve found you can’t get good ai stuff without paying something. The free stuff never meets expectations and I gather it won’t for many years to come unless it becomes commonplace.
Soon your sheri mac question will be tokenized and you will have a monthly limit you won’t be able to use any apps on your phone without them being tokenized.
The images that Kyle posted no offence intended are pretty mediocre and could never be used in any industry seeking anything resembling quality. The image with the rain looks so fake how could one use that and the image with the salt flats just looks wrong. To me the images look so wrong in many ways that a good art director would spot and expel them instantly. First their geometric composition is wrong or in poor taste. The background on the one image makes no sense in that what domestic area wants a racetrack going past it. It’s all wrong it bears no sense to reality but fools one through texturing which is at best second rate. In blender you can do superior renderings to this in seconds using an hdri and a good model. The rain looks like pieces of plastic, the motion blur is just too false to even look at. There is nothing good about this in my view except it kind of gets one in the right direction like a sketch but as artful rendering it frankly falls short.
It takes a human to give things artistic quality. I’ve watched ai adds and movies and the camera angles and everything that makes cinema great is never achieved because ai lacks the reason behind our choices. Humans make illogical choices that seem bad but in the end turn out to be more logical than logic.
Bad art and work has always existed and will as time goes on only increase. I came from a world where we did our own illustrations and plans by hand. But we still were not as good as the past artists because we didn’t and never will take the time to be great like they did and because latter we accepted shortcuts like computers that actually belittled our capabilities. Think of how the projector influenced commercial art and photography and you’ll understand why we are not as good as the past artists who sketched everything and didn’t trace from photos or project images. Why is David or Durer or Rembrandt or Leyendecker so wonderful is that they didn’t take shortcuts to express themselves nor faltered in their study or training. Why are great architects works of the past still better and more meaningful than most computer produced architecture of our day. The only true great art using computers is industrial design where math and form join but that is not done by ai that is done by visionaries.
Drawing is the probity of art and is the only way you can express yourself because it is a direct link with the non verbal part of our brains, it’s a link to our primitive past when the fire of Prometheus illuminated our thoughts and projected them out from the lamp of our imagination into reality. Drawing is the greatest invention of humankind. Each of you are unique and no one can match the other, we all have a unique gift waiting to be discovered that only each of us can reveal.
RM
beautifully said
LeMans, Spa, Indianapolis, any temporary city racetrack… probably many other tracks old and new were/are close to houses
you’re forgetting that Rembrandt, Durer, Titian and other Renaissance artists and designers rarely worked alone and owned workshops where they “employed” people to do the work. one obvious benefit of their “study” and “training” was that it helped them produce more realistic images in less time with less cost… hmm… that sounds awfully familiar.
seems like you (and most people) would be onboard with the representational paradigm of the ancient egyptians: draw/sculpt people in the same way for 1000 years all in the same proportional system. if that was the case, we would all be making byzantine mosaics and our forum would have a lot more gold and illuminated lettering.
EDIT: Durer did take shortcuts with his invention of a perspective machine. he was also a prolific printmaker/engraver (the one that makes the drawing that people use to carve the wood/stone/copper) - in his contemporary time, printing images must have been the biggest shortcut available lol
i do agree that many ai images are still in the uncanny valley, but that will eventually be resolved just as printmaking evolved. the bigger picture (npi) is how ai image-making seems to be one the first instances of a “thing” that requires humans to be electrate.
much like the adoption of the printing press required everyone to become literate (over the course of centuries), so too ai tools/ultra-ubiquitous computation will require us (well, probably my kids/grandkids) to become electrate. they are images for now, but who knows how information will be communicated once electracy reaches the level of significance as literacy (the scene in the Matrix where Neo absorbs oodles of kung-fu skills is in my mind where somehow code creates the memories and skills in our minds by moving neurons. personally, it would be nice to have the knowledge of a master Mercedes mechanic so I could completely repair my SL55).
there’s about ~500 years between printing and ubiquitous computation. in that time span, global literacy increased from “a small fraction” to “82%” (qualitatively, you likely have an intuitive feeling about how many more people could read in 2000 compared to 1500, but people like numbers). how long until humanity is 82% electrate?
Its good to have a healthy constructive discussion around new technology. As Kyle said like it or not its here to stay…..for now. And like all new technologies it will go through its honeymoon phase where everyone is excited to use it but eventually its shortcomings become more evident over time. And those arent always the look or the quality but the experience each person has with it. I dont begrudge anyone wanting to use AI or not use it. I think you choose the tools that suit you while you still have a choice in the matter. When CGI came onto the scene in the mid 90s a similar thing happened and suddenly practical effects starting dying out and the companies who did them along with it. For a time we saw all CG effects then slowly a mix of practical and digital. These days Im seeing a lot more practical being used but with a better balance with digital as we have learned to use the tools effectively rather than just because they are new. When it comes to AI for me as Ive said in an earlier post its the environmental and social impact it will have that concerns me. CGI didnt require massive data centers or consumption of copyrighted work to sustain it. We are in an uncanny valley with AI and the implications of that are far more dangerous than CGI was when it came onto the scene. CGI was primarily a visual tool used in entertainment and promotion. AI is being integrated into every aspect of our lives whether we like it or not and thats the major distinction it has from previous technology booms. People have made comparisons to previous industrial leaps forward but none have been as far reaching and impactful as AI is and will continue to be. We must be cautious about how much we place our dependency on AI because its “better/faster/ and more profitable”. What happens when it simply replaces you regardless of how much faster or more profitable you have become. Eventually we become the weakest link in the chain.
Can Runchat AI take a 3D model of say a building or limestone stair and and spit out a rendering if it’s also given photos of the stone slabs the project are to be made with….? I’m asked all the time to do renderings like this in the stone business but I have no time to learn rendering.
Brian
yes. from what I understand you can supply a model, a reference image and a detailed prompt and generate what you are describing.
I’m meeting with the run chat folks in an hour and hope to learn more.
well said and spoken with battle hardened experience.
thanks for you thoughts.
I have found that a good use of AI also. I have not been as successful having it be truly tillable, but it is a great start to create something tillable.
I live in Daytona. Racetrack noise WHENEVER and WHATEVER they are running including the 24 hour. Some people live next to airports! Everything I am going to state is from my point of reference, hopefully you can relate from your point of reference. It appears to be so on this page else I would not comment.
What I am saying does not make anything correct, nor does it become acceptable. However, moving next to a racetrack or airport - as examples - then after the fact complaining about the noise or traffic is really, or should really, be considered wasted time.
Choices can shape our lives and the outcome - even the ones made for us by others. We must do our best to live our best no matter.
When I was very young I worked in a Steel Plant - long gone to the changes over time. I learned many trades, have more than twenty certifications and am instructor certified in a few. Only I found that my management and abilities were severely limited by the fact I had no full blown degree. At forty years of age I corrected that. Careful when speaking about change, it is coming, it is inevitable. Flexibility is necessary.
What does this have to do with AI and what do I think of it? I have no opinion, yet. I do feel that all of you bring valid arguments as to the course of humanity, art, considerations of society. I am old. I might not have agreed with myself fifty or sixty years ago. My preference has become to keep an open mind unless I know it is fundamentally wrong or in someway harmful to anyone. It is an evolving landscape. Always has been - took me a while to realize that we do not individually control the way of the world. We do have some control over our immediate environment - not speaking about carbon or other issues as we must remember that words can be interpreted in many ways with differing outcomes which may even slight some - but what I do refer to is our, I’ll say our individual circles of influence. I try to focus more on those issues, hopefully to make things better for others and myself, knowing that the only way that can be truly gauged is for the benefit of everyone - the Golden Rule is still the most relevant when deciding course of action. I have seen some interesting and satisfying designs followed by AI generated movies on this site - thinking about how entertaining Lee Rosario’s Icona Vivace Concept post was [ Icona Vivace Concept ] - to me.
So what has all of that crap out of my keyboard actually impose upon anyone - except myself? Thought provoking is the best of all tools. You all have used that particular tool well.
I was thrilled to read this topic only to see that it was more about us - my opinion - than the tools. Yes, AI is a tool and nothing more unless it is allowed to become more. The thought expressed is cherished. To see so many speak without shouting, arguing but accepting and thinking what each other is trying to express is something AI can not do. Cheers to all 
and everyone. Isn;t it great that everyone, unconditionally, is allowed to think for themselves. Agree to disagree is still an option.
Disclaimer: Not responsible for speeling or grametical errors - I went to public school.
PS: great forum 
This a really interesting conversation.
I am not against AI, nor do I think it is the greatest thing since sliced bread. It has its place.
It never, ever has given me precisely what I wanted. No matter how much I fine tuned the prompts. Unless it was something sooooo simplistic I could have drawn it faster myself.
l get what Kyle is saying in regards to his business, he has to satisfy customers to make money. He is not in the business of earning a living via self gratification. Translation, if it makes the customer happy, it makes Kyle happy.
A lot of us are not in that same business. We fiddle, fiddle, and fiddle around trying to get AI to create what WE want. It never does. For a couple of good reasons:
First, it is not us. Plain and simple. Never will be.
Secondly, What it produces is “learned” through countless sources. It is an amalgamation of other peoples work. Siphoned through algorithms. It is actually not “intelligent”. That so called “intelligence” is like someone claiming they are a genius because they can look up Einstein’s formulae and repeat them like a parrot.
That said, I use it for some things. Like removing backgrounds in photos. Works most of the time. It does write some computer code correctly. Of course that is because it finds the algorithm someone else actually wrote.
My overall current impression is that it is useful for mundane things, way over hyped at the level most of us have access too, and can be a huge waste of time as we try to make it conform to our wishes.
And no, for whoever posted up there that there are homes right by race tracks. There are no homes right alongside race tracks anywhere. Maybe somewhere in Europe, but I have never seen it.
@3dsynergy I would disagree with the "not being free” argument since a lot of the cutting edge progress made in this area has (surprisingly) been free and/or open-source. At least for a good while, the absolute best AI results you could get came from open-source software which get new updates and features on a weekly basis.
It’s mind-blowing what you can achieve with StableDiffusion and ComfyUI. Sure it takes a lot longer to learn, but I’m very impressed with the fact that it’s possible to compete with the “big guys” when it comes to image generation.
And as Kyle said, it really depends on the industry and the type of work you’re doing. I’m far from being the most experienced person out there, but in many places, people just want to get a better representation of what they’re seeing. Sometimes I’m asked to make a render in 10-20 minutes because someone is on a meeting with a client and they want to show something fast. Or they can’t afford to spend too much time on 3D renders because there are more important things.
In cases like these, you’re looking for any and all shortcuts that will get the job done. If AI can create a photo-realistic result with perfect lighting/materials/textures/color matching/etc. in about 10 seconds, it’s not even possible to justify spending 1 hour on a single image.
I don’t use AI in any of my personal projects because I do them for the pleasure of learning. Obviously using AI for art opens up another can of worms but from a purely functional point of view, it’s very hard - if not impossible - to deny its power and usefulness.
I’m one of the Runchat creators, thanks for prompting the thought provoking discussion Kyle.
I agree with a lot of the concerns raised around the use of AI within creative practices. There is an intrinsic joy to learning how things work and mastering complex technical processes. Replacing these with a button click that skips straight to the finished product can feel hollow. I also share some cynicism towards the frenzied data center buildouts and their environmental implications. Most generative AI products incentivise creating meaningless content, and this is inherently wasteful. There is also a risk that models that can manipulate images with natural language will be able to automate or centralize a large chunk of the creative economy, and this will negatively impact many people who depended on this kind of work for their livelihoods unless they are able to adapt quickly.
We are admittedly walking something of a tightrope in building Runchat. We don’t want to contribute to the slopocene, but we do want to explore what we think is a chasm between what multi-modal language models are already capable of and how people are using them creatively. We don’t really want to replace technical skill and mastery, we would prefer to give creatives a toolkit that helps them explore and incorporate these new gen ai models to augment and extend existing design processes. This might help designers work faster, but we’re really more interested in whether or not we can use gen ai to make new kinds of art, or to learn new skills, or to express ourselves in new ways. If this was the case then we might think differently about their cost vs benefit.
Agreed with this as a mission statement. Thanks for Chiming in and your work in this space.
(for Folks who don’t know Gwyll, he is one of the founders of Fologram and is a pretty awesome dude. )
check out fologram here. Fologram
What I find interesting is AI was pitched as a tool to “free people up” so we can do the things we enjoy while handing off repetitive tasks to AI . What we have gotten (in the creative field) mostly is AI doing the part we enjoy while still leaving the tasks of what I call tool set up for us to do. For example, I would be totally on board with an AI in Grasshopper that allows me to tell it what I want to do and AI creates the definition for me to start playing with. As Im sure Kyle will agree with me I do not have a grasshopper brain to work with. Maybe that option exists already. Any software that requires node set up for materials not just texture creation but the final material to me seems like a likely candidate for AI. Tell it what material you are after and it creates a starting point . If any of you have spent any time in Blender you know what Im talking about when it comes to material nodes. They can be overwhelming. Having the option is not a bad thing per say . But when it becomes the only option that will concern me. Balance is key. This of course still doesnt override my concerns about the environmental , financial and social impact its having. But rather than it producing the images why isnt it allowing us to stream line the set up process to create the images but still maintain control. AI is an all or nothing system that hits the target once in a while but mostly doesnt and still requires post work. Let it help me set up for rendering but let me have the final say on the look and I hit render not AI randomly trying to find what I want. When it starts doing what it was advertised to do ( streamline the process) then I might be interested. For now its just throwing darts at a moving target and hoping one of them lands on a bullseye.
made with Rhino>runchat>fresco>photoshop
I controlled every aspect of this image, but I got a big productivity boost from the tools (rhino included)
I find this process pretty freeing so far. As a commercial artist using art to generate commerce to make a living, It’s been a long time since I did art for “fun”, so IMO…this is freeing me up to do fun stuff that I enjoy doing. 
