Something scary: AI judged photo contest results

[myotis]

These are being judged by an "AI judge" in real time with a rolling top twenty being displayed. First prize €500, and all you need to do is upload your files.

When I first looked, I thought they were nearly all awful, On a second look, the pictures have changed, and to my eyes, there are several that are OK, but the majority....

AestheticsCompetition

competition.excire.com

Excire who are behind this competition use AI to keyword photographs.

The blurb that goes with the competition says:

"The juror: our AI

You don't show your photographs to a large jury, but only to a single juror: Excire's artificial intelligence.

No personal preferences or tendencies limit its judgment.

Excire's artificial intelligence objectively evaluates your photographs.

To do this, it has learned from hundreds of thousands of photos what distinguishes good photos from bad ones.

This sample data was curated by numerous professional and ambitious amateur photographers and thus perfectly reflects aspects of image quality and aesthetics."

 

[Ed Sutton]

Grim.

 

[swanseamale47]

So photos created by a computer are judged by a computer. I wonder what the computer will do with the £500, buy an owner with photographic talent maybe....

 

[LongLensPhotography]

Maybe it will buy more storage or another GPU for itself

 

[Faldrax]

In the T&C's is the line

This also applies to photos that are obviously AI-generated

It's not 100% clear exactly what this applies to, it's in a paragraph that is about submitting original RAW files, and including certain exif.
So it appears you can submit AI images, but only it they don't look AI, and you have an original....????

 

[myotis]

In the T&C's is the line

It's not 100% clear exactly what this applies to, it's in a paragraph that is about submitting original RAW files, and including certain exif.
So it appears you can submit AI images, but only it they don't look AI, and you have an original....????

By the button to submit entries it says [my bold]:

• According to the terms and conditions, we reserve the right to remove obviously AI-generated images
• One submission per competition day is possible
• The images are not used for AI training purposes

The T&C s say "You may only submit photos that you have taken yourself and to which you retain all copyrights and rights of use."

edit:: to add to this the T&Cs also say (as you allude to) which seems to make it clear that AI generated images aren't accepted (original camera settings and capture date)

"Participants must be able to provide the unedited original files (RAW/original file) of their submitted photographs for authenticity checks upon request. Submitted photographs must contain EXIF data that reflects the original camera settings and capture date"


I would read "taken" to exclude any AI generated images, And they are just adding a rider (one of many) that thy still reserve the right to exclude those that are obviously trying send in pictures they didn't take.

 

[myotis]

So photos created by a computer are judged by a computer. I wonder what the computer will do with the £500, buy an owner with photographic talent maybe....

I don't think they have been created by a computer (Many of the AI generated images I've seen look a lot more convincing that some of these do), I think they have just been horribly over-edited by the photographer.

The worrying thing is that the AI, after analysing millions of photographs and learned lots or rules on what makes a good photograph, should come up with this selection.

Mind, it is still going to be constrained by the photographs submitted. These aren't necessarily what it thinks are "good" photographs, just the best out of what has been submitted.

 

[taxboy]

Ultimately photography is an art form (combined with a technical element). As such it should generate an emotional response in the viewer, which is why as humans you get a different answer if you you showed different people the same images.
I understand AI removes the subjective review elements together with bias - conscious or otherwise; you then create a situation where to win you effectively need to crack the algorithm and create imagery by using a tick list.
I've been put off the competitive element of photography some years ago from the camera club scene. However I do recognise that it appeals to many. It would be interesting to see what AI thinks of Ansel Adams Vs Cartier Bresson Vs .....

 

[Clive K]

The photos chosen are from a very narrow gendre. They are all like those inspirational posters that you see in executive's offices. Has AI been programmed by National Geographic? Where are the candid street photos of fat people staring into their phones, the black & white derelict canal scenes, the obviously staged wildlife images?

If this goes on people will stop buying Leicas!

 

[myotis]

Ultimately photography is an art form (combined with a technical element). As such it should generate an emotional response in the viewer, which is why as humans you get a different answer if you you showed different people the same images.
I understand AI removes the subjective review elements together with bias - conscious or otherwise; you then create a situation where to win you effectively need to crack the algorithm and create imagery by using a tick list.
I've been put off the competitive element of photography some years ago from the camera club scene. However I do recognise that it appeals to many. It would be interesting to see what AI thinks of Ansel Adams Vs Cartier Bresson Vs .....

Actually, I'm not sure you do remove the subjective/biased elements.

Someone has to have programmed the AI in the first place, Someone has to have decided where they got the training photographs from. Someone has to decide what rules the AI uses.

I've seen some AI evaluations of photographs, which have been scarily good in terms assessing the picture and advising on how it could be improved (including step by step instructions for photoshop),

But they were also.a very naive and shallow assessment, which while possibly raising some pointers for the photographer to think about, was obviously lacking any sort of sophisticated human insight. ,

Additionally, even today there are a lot of photographs held/archived in museums that are almost certainly not in their training data. Also just as another constraint I, like others have opted out of allowing Facebook to use my photographs to train AI, And some organisation are trying to stop the AI companies using their data.

What you maybe get is consistent subjectivity and bias, but as I understand it. some randomness is built into AI to ensure it doesn't give the same answer every time.

I imagine that both Ansel Adams and Cartier Bresson pictures will have used in the training and broken down in a way that the AI can assess. For example Bresson pictures are often good examples of Dynamic Symmetry which I am guessing AI could easily score by comparing it to "ideal" dynamic symmetry patterns. But I'm just guessing, I don't really have any idea how it works. That's maybe no entirely true, as I know a little bit about building Bayesian Probability Models, which loosely do a similar thing.

 

[myotis]

The photos chosen are from a very narrow gendre. They are all like those insirational posters that you see in executive's offices. Has AI been programmed by National Geographic? Where are the candid street photos of fat people staring into their phones, the black & white derelict canal scenes, the obviously staged wildlife images?

If this goes on people will stop buying Leicas!

It's difficult to know if this is down to the photographs used for training the AI or an artefact of the people entering the competition. I'm assuming the photographs entered will reflect the people who use Excire, as I've only seen it advertised in Excire mailing list

Maybe it's easier for landscapes to match the "rules" that the AI is using. i.e clearly identified subject , clear leading lines. interesting foreground.

The Photographs I've seen assessed by AI, were picking up the lack of these things as being areas that needed improved.

 

[Harlequin565]

No personal preferences or tendencies limit its judgment.

Well it's not a competition about art then.

In a random thought from an old fart, I don't think this will prevail wholesale across photography. Musicians are worried too, but look at how (arguably) punk changed the music landscape. Some people were bored with the sanitised crap they were being fed and punk (and what came after it) proved that whilst the mainstream might continue, innovative and disruptive work will still have an audience.

It's like any judged competition. If you know what the judge is looking for you can win, regardless of talent.

 

[ancient_mariner]

Well it's not a competition about art then.

In a random thought from an old fart, I don't think this will prevail wholesale across photography. Musicians are worried too, but look at how (arguably) punk changed the music landscape. Some people were bored with the sanitised crap they were being fed and punk (and what came after it) proved that whilst the mainstream might continue, innovative and disruptive work will still have an audience.

It's like any judged competition. If you know what the judge is looking for you can win, regardless of talent.

This raises some interesting thoughts. To me, as a teenager at the time when punk emerged, the UK music world threw away the good music and went running after crap, arguably taking a decade to recover. Likewise some of the more 'disruprive' art has quite literally been faecal, and while challenging in terms of content seemed to offer little more than shouting profanities in a crowd.

If AI assessment would prevent these things from being successful, would that be good or bad? I am genuinely quite torn, because I see human creativity as being a wonderful thing, but sometimes what it creates is degrading rather than enhancing. At the same time, people are shallow and run after whatever the latest thing is that everyone is running after just because it's new, without placing a value on work produced with insight, skill and wisdom over time.

Or perhaps both can co-exist alongside each other?

 

[myotis]

Well it's not a competition about art then.

I'm certainly totally unconvinced that it has any merit

In a random thought from an old fart, I don't think this will prevail wholesale across photography. Musicians are worried too, but look at how (arguably) punk changed the music landscape. Some people were bored with the sanitised crap they were being fed and punk (and what came after it) proved that whilst the mainstream might continue, innovative and disruptive work will still have an audience.

I hope it doesn't prevail, but I fear, that a well as some very good uses of AI, we are going to see a lot of completely inappropriate uses

It's like any judged competition. If you know what the judge is looking for you can win, regardless of talent.

Is it really that easy

 

[myotis]

This raises some interesting thoughts. To me, as a teenager at the time when punk emerged, the UK music world threw away the good music and went running after crap, arguably taking a decade to recover. Likewise some of the more 'disruprive' art has quite literally been faecal, and while challenging in terms of content seemed to offer little more than shouting profanities in a crowd.

If AI assessment would prevent these things from being successful, would that be good or bad? I am genuinely quite torn, because I see human creativity as being a wonderful thing, but sometimes what it creates is degrading rather than enhancing. At the same time, people are shallow and run after whatever the latest thing is that everyone is running after just because it's new, without placing a value on work produced with insight, skill and wisdom over time.

Or perhaps both can co-exist alongside each other?

In principle I think this would be bad. I don't want the people with the power and the money to build AI programs deciding what is good or bad for society.

Things are bad enough with our current news outlets ad social media, let alone adding AI, which some people seem to believe makes objective and unbiased decisions.

I would be interested to know how history now looks back on the punk scene and what it has meant for the music industry and society. As it so happens the stuff I'm currently reading (on photography) suggests that punk was only part of a post-modernist movements that reflected a range of changes in society.

 

[ancient_mariner]

The site doesn't work for me - no images visible, just a spinning circle.

 

[myotis]

The site doesn't work for me - no images visible, just a spinning circle.

I've just clicked on the link in my post and it worked. It was a bit slow to load, but everything is slow for me to load.

 

[Ed Sutton]

To me, as a teenager at the time when punk emerged, the UK music world threw away the good music and went running after crap, arguably taking a decade to recover.

Punk didn't stop me liking the stuff I'd been listening to before, it gave me something else to listen to. Sure there was a load of crap in punk, but there was a load of crap before punk. Different crap!

Punk democratised 'pop' music. "It's easy, it's cheap, go and do it!"

 

[ancient_mariner]

Punk didn't stop me liking the stuff I'd been listening to before, it gave me something else to listen to. Sure there was a load of crap in punk, but there was a load of crap before punk. Different crap!

Punk democratised 'pop' music. "It's easy, it's cheap, go and do it!"

I was a young guitar player, but nobody wanted anyone actually trying to play back then - just pick 3 chords and thrash away. It made life miserable if you aspired to a bit more than just an angry noise. A bit like telling young photographers that the only kind of valid picture is a street shot of someone on their phone.

In retrospect if I'd been 5 or 10 years older it would have seemed a bit different from a musicians POV, but experience takes time.

 

[Ed Sutton]

I was a young guitar player, but nobody wanted anyone actually trying to play back then - just pick 3 chords and thrash away. It made life miserable if you aspired to a bit more than just an angry noise. A bit like telling young photographers that the only kind of valid picture is a street shot of someone on their phone.

In retrospect if I'd been 5 or 10 years older it would have seemed a bit different from a musicians POV, but experience takes time.

Not everyone wants to be a virtuoso. As Lou Reed said, three chords and you're into jazz.

I got my first guitar in 1976 because I wanted to solo like Hendrix or Fripp. I soon realised that was too much work for me but I could play the solo from Boredom.

Digital cameras are the photographic equivalent of punk. They enable people with ideas to produce work without the drudgery of developing technical proficiency. Maybe will prove to be AI the latest enabler in image making?

 

[ancient_mariner]

Digital cameras are the photographic equivalent of punk. They enable people with ideas to produce work without the drudgery of developing technical proficiency. Maybe will prove to be AI the latest enabler in image making?

Not one bit.

Digital cameras are like synths. There's no special hands-on skill required like for guitar or drums to play your digital sample, though skills are still required to produce a good result. We have the equivalent of protools and autotune to fix your results. I can't find a good simile for the click track, but I'm sure it's there.

The nearest photo equivalent to punk I can think of is Polaroid, where anyone can do it, but it always sucks.

 

[Garry Edwards]

My guess is that this is the interesting bit (my bold) and isn't true

By the button to submit entries it says [my bold]:

• According to the terms and conditions, we reserve the right to remove obviously AI-generated images
• One submission per competition day is possible
• The images are not used for AI training purposes

The T&C s say "You may only submit photos that you have taken yourself and to which you retain all copyrights and rights of use."

There must be countless millions of photos that the programme can learn from, but they are of very mixed quality (whatever quality is) so I'm guessing that what they are trying to do is to teach their software what a "good" photo actually is, by getting "good" photographers to supply them with "good" photos.
Maybe their willingness to accept raw images, with full image data/technical info is another indication of their true intentions.

My thinking process is that although AI can be a very useful tool (I use it quite a lot to tidy up my written content, specifying things like grammar, reader educational parameters, researching and including data etc) it often comes up with very wrong info, gleaned from some website or other. An example, from an AI-generated article on studio lighting, included a statement that professional studios may have up to 4 lights, one in each corner, pointing towards the subject. In other words, AI includes nonsense as well as useful content.

 

[sirch]

To do this, it has learned from hundreds of thousands of photos what distinguishes good photos from bad ones.

And most public AIs like this are trained on what is published on the internet, this AI has no doubt learned which photos get most "likes" and and is then doing a very simple evaluation of entries submitted against most liked. To the discerning the output looks bad but they do look like the kind of photos that get a lot of likes from humans on Instagram.

Punk democratised 'pop' music. "It's easy, it's cheap, go and do it!"

Similarly the internet democratised publishing and now anyone can publish anything for free, which leads to photos like these being popular but I think we need to go a little deeper and ask why this style of photo has become popular?

 

[AndrewFlannigan]

...but I think we need to go a little deeper and ask why this style of photo has become popular?

I wonder if that's because some people write robot scripts, to give "likes" to their pictures and other posts ?

 

[myotis]

My guess is that this is the interesting bit (my bold) and isn't true

There must be countless millions of photos that the programme can learn from, but they are of very mixed quality (whatever quality is) so I'm guessing that what they are trying to do is to teach their software what a "good" photo actually is, by getting "good" photographers to supply them with "good" photos.
Maybe their willingness to accept raw images, with full image data/technical info is another indication of their true intentions.

They aren't "willing" to accept raw files, they are saying that the photographer must be able to provide the raw file to prove its authenticity, presumably before being eligible for a prize. They have to be submitted as a JPEG or PNG.

In terms of learning what a good photograph is, I've already posted that AI (most recent ChatGPT paid for version) can give a fairly sophisticated (but at the same time rather naive) assessment of photographs where it's obvious it has learnt rules of composition, including things like using colour and contrast ,something about the history of photographs where it can advise on which photographers you might want to study, and something about how to use photoshop. You get a few pages of reasonable comment.

There is a big difference between AI that has trained itself by scraping the Internet, and AI that has been trained by someone who is an expert in the subject, I think

There is an example here of getting assistance to use Capture One from regular ChatGPT, and expert trained ChatGPT

Capture One AI Assistant – Ask Expert Bot

Alex here, and today I'm excited to present my latest development: Capture One Expert Bot. This is your personal Capture One assistant, capable of answering professional questions and sharing the best Capture One tricks and image editing techniques. I have trained this bot on hundreds of my...

alexonraw.com

I think this is just marketing exercise to promote their AI driven Database software.

My thinking process is that although AI can be a very useful tool (I use it quite a lot to tidy up my written content, specifying things like grammar, reader educational parameters, researching and including data etc) it often comes up with very wrong info, gleaned from some website or other. An example, from an AI-generated article on studio lighting, included a statement that professional studios may have up to 4 lights, one in each corner, pointing towards the subject. In other words, AI includes nonsense as well as useful content.

That's pretty well what I said in my earlier post on the AI assessments of photographs I had seen. There was enough good information to prompt some thinking but without the insight that you would hope get from a critique by an experienced human being. But used well, it can be a useful tool

However, you raise a good point that a lot of the AI generated stuff is terrible, and needs human intervention to sort out the useful from the nonsense. I'm sure estate agents are now getting their marketing spiel written by AI as it's much worse than it used to be.

 

[Harlequin565]

Digital cameras are like synths. There's no special hands-on skill required like for guitar or drums to play your digital sample, though skills are still required to produce a good result. We have the equivalent of protools and autotune to fix your results. I can't find a good simile for the click track, but I'm sure it's there.

Interesting... My first thought was that I struggle greatly with synths to get anything out of them, but I can play guitar & bass (and operate a drum machine) to the point where my talent is the limiting factor. The problem with synths is the complex software that needs to be learned to make it work. I can play piano, but I have given up with synths.

Some photography has become more about computer work and less about being outside with a camera - which is fine for those that like it. Maybe this competition is more of a computer skills test rather than a photographic one (or just a big ad as Graham alluded).

I think we need to go a little deeper and ask why this style of photo has become popular?

As an uninformed Herbert, my guess is that it's "shiny". Even photographers love shiny.

 

[AndrewFlannigan]

In terms of learning what a good photograph is, ...

This just takes us back to "who defines a 'good' photograph and how do they reach that conclusion?".

Oxford University gives six things that it claims "makes a good picture"...

• are sharp and in focus
• are well lit
• have interesting angles
• have a splash of strong colour
• have something to draw you in
• show people with lively and engaged expressions

You'll note that only the first (and possibly the fourth) are objective. All the others are clearly subjective.

Under "What should I avoid?" they offer...

• have no particular point of interest to catch the eye
• are just dull line-ups of people include distracting backgrounds
• show people with unfortunate expressions
• include inappropriate clothing

Even worse, now, these are all highly subjective.

So, if even one of our "centres of exellence" can only, in effect, say "a good picture is what I say is a good picture", how on earth can a programme identify such a chimera?

Images – what makes a good picture?

Advice and guidance for selecting and displaying the best image

communications.admin.ox.ac.uk

 

[myotis]

And most public AIs like this are trained on what is published on the internet, this AI has no doubt learned which photos get most "likes" and and is then doing a very simple evaluation of entries submitted against most liked. To the discerning the output looks bad but they do look like the kind of photos that get a lot of likes from humans on Instagram.

I'm not sure if it's necessarily as simple as that (though it might be in this case), but I have no experience of instagram.

I've mentioned in other posts how AI can generate relatively sophisticated assessments of photographs, and while I don't "know" , I suspect you could fairly easily choose to train your AI bot on Ansel Adams photographs or Gary Winogrand photos etc and then use a Landscape AI bot trained on Ansel Adams pictures to assess a Landscape competition.

Personally, I still find that a horrifying prospects, but it still maybe one step up from basing it on likes.

There was a paper published some time ago (I can't remember the details) that looked into why people liked some photographs over others, As I remember, there were specific attributes that triggered liking some pictures over others, and the stronger the triggers the more we liked the picture.

I suspect that because the likes of instagram encourages devoting a millisecond of study, before moving onto the next picture, So people respond to these triggers, give it a like, and move on, before it makes them sick.

Things may have improved but my memories of supermarket ready made curries are of really enjoying the first bite, but well before finishing, wishing I'd never started.

 

[myotis]

This just takes us back to "who defines a 'good' photograph and how do they reach that conclusion?".

Oxford University gives six things that it claims "makes a good picture"...

• are sharp and in focus
• are well lit
• have interesting angles
• have a splash of strong colour
• have something to draw you in
• show people with lively and engaged expressions

You'll note that only the first (and possibly the fourth) are objective. All the others are clearly subjective.

Under "What should I avoid?" they offer...

• have no particular point of interest to catch the eye
• are just dull line-ups of people include distracting backgrounds
• show people with unfortunate expressions
• include inappropriate clothing

Even worse, now, these are all highly subjective.

So, if even one of our "centres of exellence" can only, in effect, say "a good picture is what I say is a good picture", how on earth can a programme identify such a chimera?

Images – what makes a good picture?

Advice and guidance for selecting and displaying the best image

communications.admin.ox.ac.uk

Fortunately, "good" is an entirely subjective term which only has meaning in relation to who, when and how it is being used,

A "good" meal
A "good" day out
A "good" film
A "good" friend
A 'good" photograph,

The latter in this case relying entirely on how Excire have decided to program/train the AI bot that is making the decisions.

 

[AndrewFlannigan]

Fortunately, "good" is an entirely subjective term which only has meaning in relation to who, when and how it is being used,

I agree that is how it should be used in this context.

Unfortunately, it is too often used as if it were an objective assessment of an image and therefor universal in application.

 

[ancient_mariner]

Interesting... My first thought was that I struggle greatly with synths to get anything out of them, but I can play guitar & bass (and operate a drum machine) to the point where my talent is the limiting factor. The problem with synths is the complex software that needs to be learned to make it work. I can play piano, but I have given up with synths.

Some photography has become more about computer work and less about being outside with a camera - which is fine for those that like it. Maybe this competition is more of a computer skills test rather than a photographic one (or just a big ad as Graham alluded).


As an uninformed Herbert, my guess is that it's "shiny". Even photographers love shiny.View attachment 434967

I used to use a guitar synth (Godin xtSA plus Roland GR33) quite a lot at one time, and occasionally people would come up to the stage trying to see where the sax player was. And so it is with digital, no need to learn all the fingering and embrasure that would be otherwise required. But you still need to be able to actually make it sound good.

However, I have a work colleague who is into his analogue synths. Claims to know nothing about music, but can gig with his kit still.

 

[Harlequin565]

This just takes us back to "who defines a 'good' photograph and how do they reach that conclusion?".

I would say "the viewer does" and "based on their experience and knowledge of photography".

Because photography is so prevalent and available today, and because "good" often equals "number of likes", the great unwashed of [insert social media] weigh in by volume rather than by experience.

I'd love to see someone write an algorithm for AI that allowed it to judge based on Halsman's rules....

(From wikipedia)

His 1961 book Halsman on the Creation of Photographic Ideas, discussed ways for photographers to produce unusual pieces of work by following six rules:

• the rule of the direct approach
• the rule of the unusual technique
• the rule of the added unusual feature
• the rule of the missing feature
• the rule of compounded features
• the rule of the literal or ideographic method

In his first rule, Halsman explains that being straightforward and plain creates a strong photograph.

To make an ordinary and uninteresting subject interesting and unusual, his second rule lists a variety of photographic techniques, including unusual lighting, unusual angle, unusual composition, etc.

The rule of the added unusual feature is an effort by the photographer to capture the audiences attention by drawing their eye to something unexpected by introducing an unusual feature or prop into the photograph. For example, the photograph of a little boy holding a hand grenade by Diane Arbus contains what Halsman would call an added unusual feature.

Halsman's fourth rule of "the missing feature" stimulates the viewer by going against his or her expectations.

The fifth rule enlists the photographer to combine the other rules to add originality to his or her photo.

Finally, Halsman's literal or ideographic method is to illustrate a message in a photograph by depicting the subject as clearly as possible.

 

[myotis]

Some photography has become more about computer work and less about being outside with a camera - which is fine for those that like it. Maybe this competition is more of a computer skills test rather than a photographic one (or just a big ad as Graham alluded).

Except, as far as I'm concerned, these are examples of how not to use a computer.

And maybe because I come from the film age I see being a photographer means being both outside with a camera and inside processing and printing.

I think that eventually, some might gravitate to working closely with a specialist printer who understands what you want from your prints, but for most of us, that probably didn't happen.

I feel the same with digital, and for me, learning computer skills is a core part of my photography. I just wish it was easier and didn't take up so much time.

Although, I'm not sure I would like to go back to needing a darkroom, I do feel nostalgic for the day when learning the mechanics of photography was easy, and you got better primarily though practice rather than the apparently never ending task of learning the mechanics of how digital cameras and processing software works.

 

[AndrewFlannigan]

....and you got better primarily though practice rather than the apparently never ending task of learning the mechanics of how digital cameras and processing software works.

Every digital camera has an automatic mode.
I seldom use anything else...

 

[myotis]

Every digital camera has an automatic mode.
I seldom use anything else...

Very good

 

[Ed Sutton]

Interesting... My first thought was that I struggle greatly with synths to get anything out of them, but I can play guitar & bass (and operate a drum machine) to the point where my talent is the limiting factor.

Ditto. Except drum machines and keyboards have always baffled me.

A digital camera is much easier to use than any of them.

But this thread has once again illustrated that some people place great store on technical ability, and some couldn't give a f*** about it. I've found that I often make my best work in a medium when I first jump in to it with next to no knowledge of what it can do.

It's kinda like the Picasso quote: " It took me four years to paint like Raphael, but a lifetime to paint like a child."

 

[ancient_mariner]

It's kinda like the Picasso quote: " It took me four years to paint like Raphael, but a lifetime to paint like a child."

There's a fascinating Picasso museum in Barcelona, worth a visit if you've not been. Having seen some of his work there, I wondered if he needed conventional skills first before he could become creative and develop his own vision. I think it was Gary Moore that once said "Dave Gilmour can say more with one note than Yngwie Malmsteen can say with 10,000, but at least they're both talking".

 

[sirch]

I wonder if that's because some people write robot scripts, to give "likes" to their pictures and other posts ?

Robots are a factor but I suspect a small-ish one. Social media algorithms (pre-and-post AI) can be considered as amplifiers from a systems POV. The reason we are at this point, where an AI thinks these are good photos is that contrasty, overly saturated images stand out in a grid, so they get views (for good or bad) and views mean they get promoted more by the algorithm which then creates a positive feedback loop.

 

[ancient_mariner]

Borrowed found on a music site just now.

 

[Ed Sutton]

View attachment 435019

Borrowed found on a music site just now.

View: https://www.youtube.com/watch?v=fhw7VBKytLE