MicrostockGroup Sponsors


Author Topic: I don't know why, but I feel happier when I sell AI generated photos than  (Read 11584 times)

0 Members and 2 Guests are viewing this topic.

« Reply #50 on: November 21, 2023, 06:34 »
+1
It took about 10 years to get 6,500 photos on Adobe Stock, but I already have 1,300 AI photos approved and 2,100 AI photos on review cue in less than 3 months.  Efficiency is astonishing.  By March or April next year, hopefully I'll have more AI photos on my portfolio than conventional photos.


MZP

« Reply #51 on: November 21, 2023, 08:20 »
+5
I think what MadMax meant to say was that doubling your portfolio doesn't necessarily mean doubling your income. Which is true there's no direct correlation between the size of your port and your income. You could double your income with just a couple of AI-gen images, just as well as you could 10x your portfolio and never see a significant improvement in sales.

On the other hand the more images you have, the more chances of making money, of course. But on the condition that you offer content that is varied and sellable. Adding 2000 images of exactly the same subject, probably won't do you any good in terms of sales. Also, adding 2000 images on a topic that the buyers are not interested in, will probably produce the same effect. On the other hand, adding 2000 images of varied content, that are beautifully done and free of errors (like missing or extra limbs) and appeal to buyers' needs, you could see your sales grow nicely and steadily.

But all of the above also apply to non-AI generated images. So there's nothing new here.

Indeed, it's way more efficient to produce AI-generated images than real photos or illustrations. So by generating with AI, you will see your income grow faster than by going the old fashioned way. Except you are not alone in this. There are a legion of other contributors that are doing exactly same thing: generate images with AI and submit at the same pace as you. So probably the competition stays the same. And this brings us back to the topic of quality and desirability of the images we produce. I think this should be our main focus even if nowadays all of us are able to produce 1000 images a week (with AI) versus 1000 images a year (which was the case before AI).
« Last Edit: November 21, 2023, 08:27 by MihaiZaharia »

« Reply #52 on: November 21, 2023, 08:43 »
+3
Another aspect is that, due to the limitations of AI, the spectrum of images that can be produced is limited. This very restricted spectrum is being divided among everyone using AI exclusively. As Mihai mentioned, this will lead to thousands of images representing the same concept/stereotype. Moreover, the market is already heavily oversaturated with stereotypes. An example of this limitation is illustrated by images depicting a person chopping vegetables, a person biking, and so on
« Last Edit: November 21, 2023, 08:46 by MadMax »

« Reply #53 on: November 21, 2023, 09:51 »
+2
Im on pace to more than double sales if I finish adding the same number of AI photos to my existing portfolio.  So , its all good.

« Reply #54 on: November 21, 2023, 18:03 »
+1
With the volume you are uploading you will make a lot more than double. You have a lot of experience and that makes a huge difference to all the newbies just starting out.

« Reply #55 on: November 21, 2023, 20:52 »
+1
With the volume you are uploading you will make a lot more than double. You have a lot of experience and that makes a huge difference to all the newbies just starting out.

Thank you and I hope so!!

« Reply #56 on: November 22, 2023, 13:49 »
0
Where are you selling your images? I don't see any sales under blvdone on SS?


« Reply #58 on: November 22, 2023, 19:24 »
+2
Where are you selling your images? I don't see any sales under blvdone on SS?

I'm selling them on the streets.  Or sometimes I go door to door to sell my AI stock photos.  It's not easy though.  Cash only.

« Reply #59 on: December 11, 2023, 01:48 »
0
Wow, that is so nice of you. I can understand how hard it is to sell photos on the streets and also door to door where you meet some friendly people and also some, you know.

« Reply #60 on: December 12, 2023, 14:28 »
+1
Update.
So, I have added 2,500 AI generated photos to my portfolio.  That's a little less than 30% of my photo portfolio on Adobe Stock right now.  I'm seeing sales increase especially in the last 2 weeks now.  I hope this trend will continue.  So far so good with AI photos.


« Reply #61 on: December 13, 2023, 20:34 »
+1
You are adding great stuff, especially fantastic people content.

I might have to reconsider my position on midjourney.
« Last Edit: December 13, 2023, 20:41 by cobalt »

« Reply #62 on: December 13, 2023, 20:53 »
0
You are adding great stuff, especially fantastic people content.

I might have to reconsider my position on midjourney.

Thanks!!  What do you mean by "position" on midjourney?  Subscription plan?

wds

« Reply #63 on: December 13, 2023, 22:18 »
0
Why do these tools seem to have problems with human hands?

« Reply #64 on: December 13, 2023, 22:43 »
+1
Why do these tools seem to have problems with human hands?

I agree.  The biggest problem in Midjourney is missing or extra fingers.  It's annoying to see good image with missing/extra fingers.  I don't submit them.
« Last Edit: December 16, 2023, 08:19 by blvdone »

« Reply #65 on: December 13, 2023, 23:52 »
0
I still havent used midjourney, For many reasons, one of them is that it puts out an easy to recognize look and I didnt want to have another midjourney port.

But I guess if you do interesting creative stuff other people dont do, it is less of an issue.

And it is really good with people.

We will see.

« Reply #66 on: December 14, 2023, 07:15 »
0
Why do these tools seem to have problems with human hands?

For the painters is not something new, the most complicated part to recreate of human body are the hands. Looks like for the AI too, lol

« Reply #67 on: December 14, 2023, 08:14 »
0
I still havent used midjourney, For many reasons, one of them is that it puts out an easy to recognize look and I didnt want to have another midjourney port.

But I guess if you do interesting creative stuff other people dont do, it is less of an issue.

And it is really good with people.

We will see.

What do you use if you don't use Midjourney?

« Reply #68 on: December 15, 2023, 02:22 »
+1
dall e, stable and firefly

but they are not as good, especially with people.

hoping firefly improves in 2024. would love to use an ethical engine as mymain workhorse.

« Last Edit: December 15, 2023, 02:39 by cobalt »

« Reply #69 on: December 15, 2023, 20:04 »
+1
dall e, stable and firefly

but they are not as good, especially with people.

hoping firefly improves in 2024. would love to use an ethical engine as mymain workhorse.

Just tried Firefly using the same prompts I used on Midjourney, but Firefly images are totally unusable.  Adobe has a lot of catching up to do on this. 

MZP

« Reply #70 on: December 16, 2023, 05:49 »
+3
In my experience:
- Stable Diffusion can be the same level of realism as Midjourney, but you need to use a good model, you can't rely on the standard ones. There are tons of free models available that give perfect results. It's free, it's private (since you generate on your own computer), and as long as you have a decent GPU, it's really fast.
- DALL-E beats Midjourney in almost every aspect (except realism). For everything other than photorealistic images, it's my first choice. And it has the huge advantage that you can ask it to adjust the generated images (no need for inpainting or using generative fill, just prompt in ChatGPT what needs to be changed). But it's painfully slow, it fails quite often, and it's limited to 50 generations per 3 hours (but given the frequent fails you get way less than this)
- I tried Firefly, but wasn't satisfied with the results so I gave up. I'm waiting for the next update to give it another go

« Reply #71 on: December 16, 2023, 07:27 »
+2

- DALL-E beats Midjourney in almost every aspect (except realism).  And it has the huge advantage that you can ask it to adjust the generated images (no need for inpainting or using generative fill, just prompt in ChatGPT what needs to be changed).

Midjourney can do that too. You can mark what part of the image you want changed and it will do that.

MZP

« Reply #72 on: December 16, 2023, 07:35 »
+2
Yes, also Stable Diffusion and Photoshop can do the same thing (inpainting and generative fill). But in DALL-E you don't need to paint / mark the area you need to change, you just tell it "make his hat red" or "make him look to the camera" :)

« Reply #73 on: January 13, 2024, 09:11 »
+1
Some people ridiculed me for using $120/month Midjourney plan to create AI images, but I got to say it's worth it so far.  I see the sales trend going up.  Now about 1/3 of my portfolio is AI generated images.  I'm running out of ideas these days to be honest.  So, I may need to switch to $60/month plan. 

« Reply #74 on: January 13, 2024, 09:28 »
+3
So, it took 10 years to get 6,500 photos on Adobe Stock, but took only 3 months to add 3,000 AI generated photos on Adobe Stock.  $120 x 3 months Midjourney cost is very well spent.  I'll make way more than $120 in AI photo sales this month.  Also, I'm lucky I got 3,000 AI photos accepted before the mass rejection trend started.


 

Related Topics

  Subject / Started by Replies Last post
33 Replies
8443 Views
Last post May 16, 2023, 14:10
by gameover
3 Replies
3882 Views
Last post September 20, 2023, 21:40
by blvdone
5 Replies
1431 Views
Last post November 21, 2023, 04:40
by Deyan Georgiev Photography
4 Replies
956 Views
Last post February 26, 2024, 13:47
by cascoly
2 Replies
700 Views
Last post March 09, 2024, 14:31
by SuperPhoto

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors