Insanely powerful AI feature integrated into Designer

[CM0]

Recently someone demonstrated a command window for Unity engine like the attached video.

Now imagine being in designer and typing something like "create 100 stars in a spiral".

ChatGPT can already take such instructions and turn them into SVG output. So the integration could connect to an AI system, augment the prompt so it is requesting the format to be SVG. Then Affinity could add the SVG result as a new layer. Instant AI integration allowing you to create complex objects designs using anything you can describe.

[v_kyr]

That would be ALL doable and basically is nothing else than an interoperation of some software tools & services. Pretty much like the much simpler Stock image exchange which APh etc. already have, or what some third party plugins offer for image manipulation/creation and exchange here. - Though if I'm even more lazy, why should I type/write in such requests at all, instead of telling/speaking with Siri & Co. what tasks to do for me (...greetings from J.A.R.V.I.S.)?

However the overall downside (cons) is, that one will get pretty "lazy and stupid" over time this way and so may forget how to do such things (the tasks) manually or programmatically oneself. Other cons are that you always need to have an internet connection in order to do these things and that services like ChatGPT may not entirely correctly understand what you realy want them to do.

Thus I would personally always first of all prefer to make such things individually on my own via scripting and the like. So in order to create some spiral starfield galaxy I personally would tend to go instead more this route here.

[CM0]

3 hours ago, v_kyr said:

 downside (cons) is, that one will get pretty "lazy and stupid" over time this way and so may forget how to do such things (the tasks) manually or

You will still need to do almost everything you do now. Integrating the objects into your scene you will still need to precisely manipulate them.

I see this type of interface more of what I want from AI. That is AI as a tool and not a machine that just creates a random end product like the image generators.

However, on a societal level, your observation may be correct if AI does everything for us. It does have troubling implications. At the same time, as a task accelerator it is incredibly useful and every application is looking how they can take advantage of that capability. There most likely will be a plugin for Adobe at some point, maybe Inkscape as well. It is likely an easy integration, but you can imagine how well this could sell a product with slick video demos.

[v_kyr]

2 minutes ago, CM0 said:

It is likely an easy integration, but you can imagine how well this could sell a product with slick video demos.

Of course and especially due to the overall actual industry AI hype, where everyone tries to jumps onto that hyped bandwaggon. Also we already saw from some other photo based tools, that some of them already make use of some AI features and also use that hype as a newly selling point.

[CM0]

Well, just after I posted this, Adobe just announced essentially this idea. Generating custom vectors, brushes, and textures

https://www.adobe.com/sensei/generative-ai/firefly.html

[debraspicher]

AI has its place. I think like anything else, it just depends. I don't feel keen to depend on a bot for my workflow. That sounds like a quick way to end up under another subscription fee. Better to create my own scripts. For brainstorming workflows or iterating through options, it's highly useful to use AI, at least once in a while. For implementing design decisions, I feel that is its weakest point. Inevitably, the same patterns and expectations that come from "easy design" will make AI solutions very predictable. In those cases, a basic script is better because it's basically the same solution but repeated... and then human touch is needed to involve actual creativity, not mindless "make an AVI of spinning logos with a meteor approaching" over and over and over... but again, even that has its purpose and I think that's fine. I've learned from listening to AI created stuff enough, that it takes far less time for that to become repetitive and to get bored, so I don't find it as threatening as some. For a small creator who is just starting out, it's fine to generate quick things to get a channel or web project up and running... I still remember way back during the beginnings of web we had ways to create fancy 3d logos (even animated) using a form generator, but it's not like we stayed dependent on it.

[debraspicher]

29 minutes ago, CM0 said:

Well, just after I posted this, Adobe just announced essentially this idea. Generating custom vectors, brushes, and textures

https://www.adobe.com/sensei/generative-ai/firefly.html

At what they cost, I would expect them to add this feature in so people can recoup. Small businesses who do their materials on their own and are time-limited.

[PaoloT]

Shame I can't post an exchange of words between Jackie Treehorn and the Dude in The Big Lebowski, here…

Paolo

 

[CM0]

1 hour ago, debraspicher said:

I've learned from listening to AI created stuff enough, that it takes far less time for that to become repetitive and to get bored

Yes, I agree. Which is why I prefer it being used as a tool vs. creating the end product. It can be very helpful as tool. Scripts will still have their place, but nothing can compete with the productivity potential. AI can even create the scripts. When used as a tool you have full creative control. That is very different than throwing random words at something like stable diffusion and pressing generate over and over again until something interesting manifests. The implementation I'm suggesting would essentially be nothing more than like smart macros.