Menu
What's new
The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Levi Strauss uses AI models now

KC_2020

KC_2020

Active member
Paul Spinnler said:
AI WILL definitely have an impact on fashion photography businesses ... I am afraid ...
Click to expand...
In the same way that self checkout took away the jobs of cashiers and robots replace warehouse workers.

There's nothing new here. Just Levi leveraging technology to cut costs and portraying it as being conscientious.

Profoto bought Styleshoots and introduced their branded e-commerce studio system. Will they replace some of the people who shoot products for $8-$12 a shot ? Yes. Though they still need models. So AI may replace both at scale but probably won't eliminate small businesses who need a dozen photos. See Etsy.
 
P

Pieter 12

Well-known member
I have not read the website--too much mumbo-jumbo. But if Levi's is using it to advertise their clothing and using 3D models of the clothes instead of the actual thing, that would be misleading advertising. A good lawyer could have a go at them, especially since they have deep pockets (pun intended).
 
P

Paul Spinnler

Well-known member
It is 720 EUR per month for unlimited content - nothing for a corporate like them! Difficult to beat via classic photographic means, even if you have automated studios, etc.

You can “render” small, medium, large and XL people in all races so the customers “identify” more with the company.
 
Last edited:
Ray Harrison

Ray Harrison

Well-known member
Hopefully people will still enjoy human made things featuring real humans. I tend to prefer them :) .

I love this: "Our business is centred on building human-like avatars". No human photographers or models. Sounds cold and clammy to me.
 
R

richardman

Well-known member
In addition, claiming this would increase diversity is very disingenuous. What they are doing is actually taking away potential jobs from creators and models from marginalized communities, and also, effectively silencing them.
 
P

Paul Spinnler

Well-known member
The reality is we are just at the beginning of the generative AI revolution and the pandemic seems have to been a turning point for high end MFD with sales breaking down for various reasons. This is now on top. Wedding photographers are probably not anytime soon rushing to get IQ4s - for many reasons including also such systems not being fit for a wedding in terms of speed, sensitivity, AF, weight, etc.

If high end image creation continues to improve there will be a future scenario where AI will be able to create higher and higher end imagery from the comfort of a sofa.

Especially landscape photography in my view is at risk. With thousands and thousands of similar images of the same spots one should at one point be able to create new variants with ease. All these Iceland pictures - or national park spring tree pictures … for example.

Stable diffusion and midjourney currently produce small images, but some artists like Tim Tadder are weaving this technology into other post production workflows to create stunning results …

This company here though is really disingenuous and it is a bit concerning to see how corporates will use this technology in marketing.
 
Last edited:
R

richardman

Well-known member
AI will change the world.

Commercial photographers - headshots, products, fashion, wedding, even school photos etc. - will become more and more "we can do it with AI". And in particular, marginalized groups will lose even more of their representation, ironically, by companies, like the one mentioned, who claim to want to increase representations (such BS)

And as Paul mentioned, even fine arts like landscapes. You want the most vivid Yosemite picture with amazing milky way and stormy weather? Sure why not. Print it to 6 feet for your living room, and delivered for $199.

Ditto with artists. Painters...

Novelists? Sure, it may not be David Gerrold yet, but romance novels, mysteries, ...

and News stories. That "new murals discovered in ancient Egypt", the photos of TFG being arrested? Yea, fact check them, and again.
The world is changing. I don't think we know where it will land yet.
 
P

Paul Spinnler

Well-known member
I think with a bit post production one already now should be able to wip up some trite Tuscany, Iceland, Scotland pics with Midjourney. You can resize with gigapixel and then blur it a bit and add grain and voila you are an artist.

One can skip the overpriced workshops directly. Stable diffusion runs quite fast on the local MacBook take just a few minutes to install.
 
P

Paul Spinnler

Well-known member
This took 30 sec to create on an M1 Max; added a Kodak LUT on it and some grain. And we are in the stone age of this. Like when basic HTML web pages were all the rage in 1995 ... this is on stable diffusion running locally without control-net extension. Quite passable, although the trees in the middle are bit boxy.

Especially abstract fine art landscapes will become a lot easier with this. All that B&W stuff with ND filters, etc.

1679826223781.jpeg
 
Ray Harrison

Ray Harrison

Well-known member
Technology shifts are obviously going to happen. But it rarely supplants the old completely. There's still a flourishing market for physical books, vinyl records, film and other "analog stuff". Well-made paper and pens are very much still with us. We humans seem to love our mechanical tools, for whatever reason :). At the end of the day, new tech often becomes additive rather than a complete replacement, I feel. When new tech hits the market, for example with all of the seeming magic that AI/ML brings, there's a flurry of everyone trying it out and it is disruptive, for sure. There are always job losses. For those who's main goal is trying to drive cost out of their business and who are OK with the results, AI-generated content is going to be interesting for them to try, I'm sure. My company is already rubbing its hands together with dreams of replacing software developers with AI. Too, the Levis move is obviously (in my mind anyway) cost-cutting and has no bearing whatsoever on diversity and inclusion.

If everyone has a 6' print of AI-generated Yosemite on their walls for $200, it becomes less of a talking point than someone who has one taken by a "named" human using optics, light, a digital or film sensor and with human judgement and finesse. "Hand made" always seems to carry value, in whatever venue, even if it takes time to adjust the market for it.
 
S

steveash

Member
Another new development in this space is AI Render for Blender 3D. A user can build a simple 3D model and then allow Stable Diffusion to fill in the gaps to produce a lifelike or creative render. Another step towards making commercial photographers unnecessary. As this is my livelihood I’m starting to wonder if I should keep screaming at the crocodile not to eat me or try and ride it’s back. It certainly isn’t going to go back in its cage.
 
D

daz7

Active member
It will take a lot of time for AI to be comparable with humans, art wise. Most AI 'artwork' is pure kitsch, really.
AI trains their model on what is available online and most online pictures online are purely amatourish - over 90% is of primary school art lessons quality.
I have not seen yet a single AI created art that I would consider great.
It is all poor taste, mediocre at best.
Maybe it will improve wtih time but then, to train it, the model would need to take in real masters only and that's difficult without serious copyright infrigement.
 
P

Paul Spinnler

Well-known member
You can train a model with your own data or data that suits you. Remember, we are 1995. This is just the very, very beginning.

SoC will integrate increasingly high-end AI engines and the choice of models will increase over time as will resolution. Right now models are trained on tiny images - only a question of time until training based on high-res repro scans of fine art will be possible.

You can also mix - augment your art with AI elements.

I think it will become more and more pervasive as time passes and maybe AI will also be integrated into cameras at one point directly. Imagine a Leica M13 which can output van gogh like pictures from the scenes it photographs, etc.

There are already online demos of stable diffusion running on Android phones and rendering images in 15 seconds or so ...
 
P

Paul Spinnler

Well-known member
steveash said:
Another new development in this space is AI Render for Blender 3D. A user can build a simple 3D model and then allow Stable Diffusion to fill in the gaps to produce a lifelike or creative render. Another step towards making commercial photographers unnecessary. As this is my livelihood I’m starting to wonder if I should keep screaming at the crocodile not to eat me or try and ride it’s back. It certainly isn’t going to go back in its cage.
Click to expand...
Thanks for this pointer - this is also a very interesting angle
 
Ray Harrison

Ray Harrison

Well-known member
steveash said:
Another new development in this space is AI Render for Blender 3D. A user can build a simple 3D model and then allow Stable Diffusion to fill in the gaps to produce a lifelike or creative render. Another step towards making commercial photographers unnecessary. As this is my livelihood I’m starting to wonder if I should keep screaming at the crocodile not to eat me or try and ride it’s back. It certainly isn’t going to go back in its cage.
Click to expand...
What has served me well in my industry of software development is to "ride the crocodile". The ability to pivot and move - to "problem solve and have a broad/deep toolset" - is supremely valuable. Flexibility is everything.

And we can still scream :) . "Rage, rage against the dying of the light."
 
You must log in or register to reply here.
Top