AI generative Fill in Affinity

[LuBre]

21 minutes ago, carl123 said:

Soon, highly skilled professionals could find themselves working for pennies just to put a meal on the table.

 

True, but that's always been the case.

I could make milions in the early 2000's by simply adding a database or a funny Flash animation to a website. I still remember selling animatefd GIFs for €300-500 (each). Anyone here rememebers embedding QuickTime VR on a website? That was an easy 1000-1500€ per video. Clients went NUTS, it was like showing them the future.

Back in the day you could make so much money with very little effort. Nowadays we still get paid very well, sure... But we also do/offer a LOT more for our clients, at the same price. Did anyone lose their job? Not at all. AI is nothing different than other tools we've had in the past. It's just yet another tool under our belt. We will adapt and we will offer better/different services while keeping similar prices.

[LuBre]

1 minute ago, LondonSquirrel said:

If you had the letters CCIE after your name, that was the key to £100,00 a year back then.

"I know a Flash expert who can do that"
"I know a database expert who can do that"
"I know a #buzzword expert who can do that"

The glorious late 90's early 00's.
 

[CM0]

15 minutes ago, LuBre said:

"I know a #buzzword expert who can do that"

The glorious late 90's early 00's.

lol, yes the buzzword economy. Nothing delivered what it promised. Usually it just delivered the exact opposite, complexity that only hurt your business.

I remember turning simple FTP batch scripts that cost nothing to execute into monsters of BPEL SOAP services just because executives wanted to promote advanced tool and business process adoption. A script that took someone a few hours to write turned into half a year project for an entire team. Madness.

[v_kyr]

1 hour ago, CM0 said:

However, their scripting API would provide the necessary integration which avoids these limitations. 3rd parties and open source communities are going to be far faster at rapid innovation and keeping up with AI.

Wait and drink tea (in my case coffee), since nobody knows what it finally will offer and what not (so how deep the integration capabilities it then offers will be, or if it will have certain restrictions).

[lphilpot]

1 hour ago, LondonSquirrel said:

Three mid range Sun Enterprise machines to do the work

We had numerous E3000s, E3500s, an E5000, a couple of E450s, two V880s, a bunch of 1 / 2 / 4u SunFire servers (201, 245, 280R, 440, etc.), a 3800, a 4800, a 6800 (IIRC), a couple of T4s (?) and more SSAs, Dilberts, T3s, D240s and Ultra 1, 2, 3 & Sun<whatever> 100 workstations than you can throw a stick at. Of course that's between 1997 and 2013 more or less. Lots of SPARCs, usually directly or indirectly running or supporting many Oracle databases and the apps that ran on them. One of the V880s was my primary (but far from only) responsibility, as it was a backup server with (SUN branded) StorageTek SuperDLT robotic libraries attached. It replaced an E3000 with ADIC DLT7k libraries.

Solaris was a good OS and the SPARC servers were bulletproof but they lived in the wrong space for my preferences (i.e., enterprise). Gimme a small unimportant Linux system anytime.

[CM0]

23 minutes ago, v_kyr said:

Wait and drink tea (in my case coffee), since nobody knows what it finally will offer and what not (so how deep the integration capabilities it then offers will be, or if it will have certain restrictions).

True, but Affinity at least has potential opportunity here by not already having an API they have complete flexibility in design, in that they can observe what is happening in the world with AI and the use cases and ensure that such integration will be possible, highly efficient and rich in capability.

[v_kyr]

1 hour ago, LondonSquirrel said:

Yeah. Around 2000, I set up the infrastructure for a web site. Three mid range Sun Enterprise machines to do the work, a pair of F5 load balancers. Sitting at the end of a 2Mbit/s leased line.

This infrastructure looks/sounds familiar to me, but with parallelized modules Sun servers. Though I think for me it was before 2000 or at least earlier than that. - We once spontaneously received an order from company management to create a distributed, multi-user email campaign system as quickly as possible, beside our usual project work. The given requirement at that time was to be able to send ~1 million emails in just one hour of time. At that time we were attached/connected to an outer high-speed network ring where parts belonged to the company. So the network speed was no problem, but generating and sending out that much emails in shortest time. Here the Suns came in  to be handy and our Unix-sysops gurus set the whole up in a balanced multi Email-Server structure, so the machines could throw out masses of Emails very quickly in parallel.

As far as I also remember, the problem back then was finding a suitable remote location for testing (i.e. where you could send en mas emails in the shortest possible time and get a counted in time result back). - At that time, many of our developers were always receiving a lot of unwanted spam emails (the spamming never stopped despite requests from us) from a certain US server. Therefore, our first unofficial test consisted of sending a few reply emails to the identified US server. As far as I remember, the US server in question ended up with ~15,000 emails and then was no longer accessible for ages (... and yes, the spamming was done for the future)! 😇

[lphilpot]

4 minutes ago, v_kyr said:

So the network speed was no problem, but generating and sending out that much emails in shortest time.

That's easy - Just install Veritas Volume Manager on a SPARC server, then pull a disk.  😁 

I still remember when a disk died one night on one of our old E3500s under VxVM control. I was oncall and ended up with more than 14k emails in my inbox the next morning, all basically saying the same thing: "<servername> has reported a failure on disk c0t1d2" (or some such). And for each of the emails (which represented a unique alert) the was theoretically a matching page. The only thing that saved me there was a built-in thing in our paging software that would auto-reject identical messages within 30 seconds of each other. Our then Lotus Notes email infrastructure apparently didn't drop any message (darn it).

...and don't even ask about when one of the process controllers (?) on your EMC SAN spontaneously reboots, forcing a failover and then an immediate failback of every volume on the fabric. At least it proved the process worked, but wow... all the "it's the end of the world!! no wait... I'm OK" alerts from the SAN and every server attached to it, multiplied by the multiple number of fiber paths / HBAs and logical disks.

Man, I'm glad I'm retired!  🙂

[VectorWhiz]

On 5/24/2023 at 2:20 PM, Lee_T said:

Hi WMax70,

There are no plans for this at present.

Lee

If I were Serif I would at least seriously consider adding AI assisted functions to the suit, particularly since AI recommender systems are capable of generating code on request, that may need a certain amount of tweaking, but most likely saves coders a lot of time and headaches, while extending the programs' possibilities significantly. It seems to me that AI assisted functions will become inevitable if they're not yet imperative already for surviving in the market segment of graphical art and design. I understand that from certain angles there may may arise discomfort or even opposition against incorporating AI based functionality, but programs operating in the design market have already opted to develop in that direction for equally good reason that may cause die hard supporters of exclusive human creativity to not use AI. The latter group may actually face the situation of dying hard if it persists in clinching to its principles in that field.

[Henric]

On 5/25/2023 at 6:57 AM, carl123 said:

Worth watching if just to see/hear how Adobe gets around any copyright issues.

I would also imagine this is scary stuff for those professionals that are currently paid a lot of money to do this manually

 

 

Work good many times.

 

At the end of the video was really odd and funny! 😆

 

What the ai-world happen here 😆

[Jeep570]

Well if AI doesn’t come into Affinity in the very near future then it is dead, a very great shame, as i moved to Affinity from Photoshop, but AI is a game changer and would save me hours on the editing process, for some shoots, particularly sports, weddings, and such 

[firstdefence]

AI will be added at some point in the not to distant future, I am sure of that, unfortunately it has to be added because professionals can no longer function effectively without it.

I wonder if users think that the Affinity AI that is an inevitable addition will be free to use as much as they like or like all AI based generative products the AI will be a per generation paid for bundle service?

As for Affinity being dead because it doesn't have a productivity feature, I doubt that very much lol!

[carl123]

On 10/28/2023 at 12:05 AM, Jeep570 said:

but AI is a game changer and would save me hours on the editing process, for some shoots, particularly sports, weddings, and such 

Once AI is mainstream you will no longer be needed

To survive you need to ensure your skills cannot be replicated by an AI

 

[alecspra]

Very interesting conversation! For me the most important task that I would love AI to take over is selecting the background or sky of an image. I found the process in Affinity slow and tedious especially if it is a complex image. It's not really a creative or enjoyable process so I don't have a problem letting AI make that selection for me as long as I can always make some final adjustments if needed. 

[Horseflesh]

On 5/26/2023 at 1:35 PM, nezumi said:

I just really dislike generative AI spitting ready artwork. Its just... wrong on many levels and sad on many more.

It's analogous to the era when digital cameras became cheap, and phone cameras began taking better pictures. The market was flooded with amateur pictures. People like me, who barely know what end of the camera to hold, were taking money out of the mouths of "real" photographers.  I've had mediocre photos in magazines, newspapers, and corporate web sites. I myself was the shoddy generative AI in this analogy, along with countless other amateurs.  We provided a product that was 1) Cheap or free 2) Available immediately 3) Good enough. Oh and 4) Often on shaky legal ground because no one gets model releases, haha. 

Generative AI art has all those traits and it's here to stay just like the mediocre free wedding photos taken by your cousin's best friend are here to stay. It has not killed the pro industry but it has changed it. You do not have to like it, but you will have to adapt to it. 

I do not blame Serif for not jumping on the generative art bandwagon though. As others have said there are plenty of day to day usability issues in the Affinity suite to address first. I'll enjoy generative AI features if and when we get them but there's a long list of other things I wish they would do first. 

The legal landscape around generative AI is also still taking shape. If I ran Serif I'd sure be worried about getting sued by a rightsholder claiming their work was used improperly for AI training. Who wants to be a legal pioneer in dangerous territory? Likely not a smaller firm like Serif. 

 

[lphilpot]

7 minutes ago, Horseflesh said:

I do not blame Serif for not jumping on the generative art bandwagon though. As others have said there are plenty of day to day usability issues in the Affinity suite to address first. I'll enjoy generative AI features if and when we get them but there's a long list of other things I wish they would do first. 

"Generative" doesn't have to mean "total creation". For what amounts to "inpainting on steroids", I'm totally OK with AI. For more intelligent selections, more accurate masking, more realistic colors, shading and tones, etc., etc., bring AI on. Can't be soon enough.

But I have no interest in totally AI-generated imagery. As a hobbyist photographer, AI won't affect my photography. I'll shoot (or not) by my on accord, not because I have to. But in a much larger sense AI will obviously have a huge impact and probably more often than not, a negative impact.

As a side note, I find it amusing that totally AI-generated images are often credited as "AI art by <firstname> <lastname> via Midjourney" (or whatever). Why even include a name? They did nothing but write a request. They couldn't predictably nor accurately change the output imagery through their request anyway.

[Horseflesh]

The request itself is a creative act, I think it counts... But how to divide up credit... We'll be arguing about that for the rest of time!

[nezumi]

If you want to stretch it then anybody who had any idea ever technically was part of creative art. That will make literally everybody an artist even if they never actually created any art. Every commissioner, every client would suddenly become artist because he had idea - therefore took part of creative process, right?

No, he didnt.

Thinking about something and actually making it is whole different story. Anybody who have ever tried to draw something will quickly notice that bringing to life, to physical world this idea, that seemed so clear in his head is in fact very vague, missing details and his hand is not drawing at all what he had in mind. It takes many years of training (hey, its me, captain Obvious) to actually make ideas reality. If somebody commissioned a piece to an artist, or wrote a prompt - he is not creator. He commissioned it precisely because was unable to create. He can have all the philosophical excuses in the world, make all the mental gymnastics in order to convince others and himself that he also is the creator. But everybody knows reality. So theres nothing to argue really. If you have paid runner to win a race you are not partly athlete Its just absurd and quite sad to see people even bringing that silliness up. If you can do iy YOURSELF you are the creator. If you need to ask somebody/something to make it for you - you are not. Simple as that.

[Horseflesh]

> If you can do iy YOURSELF you are the creator.

Then why do we use tools like the Affinity suite? I certainly cannot place every pixel in a document. Tools fill in the gaps. 

A film director isn't a creator?

You don't have to like or value generative works but saying that the output lacks all creativity seems to be going too far.

In any case it doesn't matter for Serif products as it isn't happening any time soon. 

[loukash]

45 minutes ago, nezumi said:

That will make literally everybody an artist even if they never actually created any art. Every commissioner, every client would suddenly become artist because he had idea - therefore took part of creative process, right?

Frankly, in the digital domain that's nothing new: Ever since some of my clients got themselves a computer with MS Office in the 1990s, they started to feel like they can call themselves "graphic designers", too. This trend only continues to grow until today. Occasionally, some of them will at least get back to me and ask for (paid) help because the print house would refuse to print their "layouts"…  

[Horseflesh]

I think it is perfectly fair and fun to argue about the nature of the creative process and the value of various contributors. 

A pointy-haired boss with a clip art CD or a generative AI tool is not Leonardo DaVinci, but they are creating something. 

That's the great thing about computers -- they give more people the ability to express themselves. Even if most of the results suck, it's a wonderful thing! 

I took an art class once where the instructor started the first lecture by talking about how one of the great masters would find natural chalk, and he used that for his sketches. Man, if I had to dig my own chalk out of the sea cliffs and learn to draw to be a Real Creator, I'd never have any creative output at all. Though by the standards of this forum, perhaps what I do has never counted as creative. 

 

 

[R C-R]

34 minutes ago, Horseflesh said:

You don't have to like or value generative works but saying that the output lacks all creativity seems to be going too far.

I am not sure anyone is saying that but to me the most important issue is determining who (or what) is contributing enough creative input to claim it as their own work, legally, morally, or otherwise.

[nezumi]

Look, if somebody wants to feel like he is creating something by simply writing what he wants thats fine. If you want to consider a "tool" something that does everything - I cant stop you. People have all sorts of fantasies about themselves, all sorts of delusions and I will not judge that. But also dont expect me to call you a creator when all you did is telling what you want. By this standards you would consider yourself a builder because you said a guy where you want to have a window and he did it for you. An electrician - when you said to a guy that you want to have a new lamp hanging from the ceiling and he did it for you. A plumber - because you have pointed to the sink thats clogged. And so on.

If you want to consider yourself a driver while sitting in the passenger seat thats cool with me. But please kindly dont expect me to join you in believing in that illusion. And by all means - have fun with your toy steering wheel making "brooom! brooom!" noises while somebody else is driving. Its important to have fun in life.

Now if you excuse me - I need to catch a plane.

I'm a pilot, you know. I told the person at the counter where I want to fly.

[thomaso]

42 minutes ago, Horseflesh said:

You don't have to like or value generative works but saying that the output lacks all creativity seems to be going too far.

Do you use "generative" on purpose? It is not identical with the discussed "artificial intelligence" for image creation or automatic image background expansion based on huge data bases of existing images.

While "generative" may mean "based on algorithmic formulas" it is used already in Affinity for instance for "Procedural Textures" in APh or in tools like the "Inpainting Brush". Accordingly "generative" design is common since many years, in 2D and 3D design and architecture. For instance …

https://www.generativedesign.org/
http://www.generative-gestaltung.de/2/#sketches
https://processing.org/books/

[R C-R]

2 minutes ago, LondonSquirrel said:

That's a good point. The director is the director, whether they take ideas from actors, extras, set designers, stunt performers, the sandwich delivery person, or whoever. Else the director credit would be 10 lines long.

But again, I have to wonder how we should decide who or what should get credit for creating how much of the output. Is there some quantitative way to do this or is it all just guesswork?