Stable Diffusion, The AI that paints your picture

Greg Taieb · · 1 min read

I was talking to you about it a few weeks ago but it's just amazing how quickly new developments are coming in the field of image generative AI.

Latest example, the release last week of the first concrete applications which rely on the Stable Diffusion model and a Google application to produce from of photo samples some pretty crazy images.

It works with products, animals and of course human beings.

All the tools to do this are free but not necessarily easy to use for ordinary mortals. As a result, very clever developers offer “turnkey” solutions to test in just a few minutes.
Only problem: it pays.

The principle:

  1. You go to the site Astria.ai or Profile Picture AI
  2. You create an account and buy credits (I'll come back to that)
  3. You load around twenty photos of yourself so that the AI can learn
  4. 1 hour later you can start typing requests that will generate images unpublished

You know me, I of course tested it and I must admit that it produced some pretty crazy results.

And yet I did not completely follow the recommendations on the images to load so I think that the AI could have been even more relevant on its results.

Of course I advise you to test it because it is amazing and it opens up crazy possibilities for the future.

On the other hand the negative point is that the developers who launched these applications “user friendly” are riding the wave and increasing their prices every day.

Astria started at $4 and at the time of writing costs $20 to test on a model. Same thing for its competitor Profile Picture AI.

To be tested urgently if you want a unique profile photo.