Learning Gimp vs. Waiting for Automated Solutions Based on Deep Learning

I am standing at the foot of the mountain known as Gimp and trying to figure out if it is worth scaling (i.e., learning to use it) or if I should wait for AI-powered automated solutions to catch up and make it irrelevant. Or, will that never happen?

Many of you have no doubt seen recent developments such as this:

A growing number of photo editing tools seem to be incorporating AI, and I wonder if learning Gimp is like going out to learn how to drive a truck. Truck drivers will be replaced by self-driving trucks eventually, although a human will be required to still ride along - at least for a while - to supervise the vehicle. Perhaps a similar fate faces human photo editors or am I being alarmist?

I see tools such as Photolemur and wonder if there is a FOSS alternative/equivalent. And, will Photolemur displace manual photo editing within 20-years?

It’ll never happen.

No matter how good the tools get, you still need to decide which results are more pleasing to your eye.


How long you want to wait? There will of course be much improvement coming, but I think photo editors will never be perfect or “finished”.


It is a common misconception that you have to learn the program fully before being able to use it effectively. I suggest you ask yourself the following question
What is it that my photograph needs so that it is a better representation of what I have in my mind?
Then check if a given program can be used to create a effect you want. Go on with this process sufficient number of times and then you have learnt the program. There is no Mountain to climb.



The “mountain known as Gimp” is, ironically, the slippery slope that got me into photography. After years of frustration, having to deal with Windows and the nonsense of paying for Norton that was ineffective, a friend who knew more about computers than me, told me to try Linux Mint using the dual boot option. Gimp 2.6 was included with the distro at that time. I liked the idea of manipulating photos, and thus, started my climb up Mt. Gimp. After a few years of Gimp tutorials my interest shifted to wanting to be able to take photos as well as manipulate them.

Analogically, I guess one can look at it from the perspective of “Why learn how to cook, when I could go to a restaraunt?”

Speaking only for myself, I’d rather know how to cook… and use a camera… and be able to form images into what pleases me.


Piling on, learning GIMP isn’t just about learning GIMP, it’s about learning the ins and outs of digital imaging. With that understanding, you can more critically consider the effects of more-automated approaches.

With regard to AI, I’m pretty sure the technologists will be able to build engines that they can then train to accommodate the expectations of a large majority of picture-takers. But, down the road, what if one of these produces an image for you that doesn’t please, what do you do? Pull out GIMP…


Indeed! As I think about teaching workshops, the first step in the endeavor is to have a feeling about a thing and then being able to visualize that thing in a photographic medium. It doesn’t have much to do with clicking buttons in software X.

Further than that, AI imaging software will be trained on popular imagery, because that’s what people like. The AI will spit out Instagram image after image. AI (currently) can’t make the image I see in my head. And when it actually can… I hope I’m dead by that time :stuck_out_tongue_winking_eye:


I thought Gimp was relatively easy to learn. Much easier than many other s/w I’ve used anyway.

AI is cool like self driving cars when your drunk or don’t want to drive, but I would rather have a sports car that I can drive myself for most of the time.