The (AI) Coca Cola Chrismas advert

myotis

myotis

Suspended / Banned
Messages
4,503
Name
Graham
Edit My Images
No
Not sure where to put this, but Tinhouse Studios has made another critique on AI and how it is affecting his photography business, which I always find interesting.

This time around, it's the Coca Cola Christmas ad that triggered the post.

I've linked to his video (12 minutes) and the Coco Cola ad.

I don't feel the same way about the ad, as so many of the commentators do, even if I understand the points they are making. For me, it's obviously fantasy from the start and I didn't find myself expecting any reality, and didn't pay any attention to the details of the content.

This is, however, because of it's obvious fantasy, in contrast if I am watching a drama (which is meant to real), I can't help noticing if a car in the background was only released after the drama was meant to have taken place, or a picture on a mantle piece appears and disappears as the drama progresses or sounds from a stills camera being used in a scene don't match the sound that this particular camera actually makes etc etc.

That aside, I feel it's an interesting perspective on the AI discussion.


View: https://www.youtube.com/watch?v=3PfWzHpT_Kw



View: https://www.youtube.com/watch?v=Yy6fByUmPuE
 
Interesting interview with the boss of Waterstones on BBC 'Today' this morning about AI books - he's totally against it ... but will stock them as long as they are upfront about it. :thinking:
 
gramps said:
Interesting interview with the boss of Waterstones on BBC 'Today' this morning about AI books - he's totally against it ... but will stock them as long as they are upfront about it. :thinking:
Click to expand...
The simplistic solution is for us to just not buy these things

But the marketing people know exactly how to manipulate the masses into buying things they don't need and, all too often, are actually bad for their health and wellbeing.
 
GreenNinja67 said:
Of course he will, he runs a business.
Click to expand...
I don't understand what this refers to: "of course he will" what?

EDIT: ah I now think you were replying to "Gramps" and not to my post, so I now understand.
 
Last edited:
gramps said:
Interesting interview with the boss of Waterstones on BBC 'Today' this morning about AI books - he's totally against it ... but will stock them as long as they are upfront about it. :thinking:
Click to expand...
The sad reality is that most books will be part written by ai from now on. It could be anything from basic structure and outline, to filler material, ideas, examples to pretty much the whole lot. It is slow and expensive work so the use of ai makes a lot of commercial sense despite all ethical concerns
 
I can remember people being dismissive about mobile phones and the world wide web but here we are. All of these technologies are disruptive, revolutionary and here to stay. As someone said "technology is the word we have for things that don't work ... yet"
 
Artifice has been used in film making since the days of silent movies e.g. Buster Keaton in the click tower scene to the car chase in Bullet with Steve McQueen where the engine/gear change noises of the Mustang he was driving were from another vehicle.

Such usage of the 'unreal' has since become computer driven & now generated. An item on BBC Tech Life (World Service) was about AI produced music and that there are currently two "AIs" for music production.

The world, based on changes to date, appears to be going through an 'AI revolution' and akin to the industrial revolution where the transition from artisanal manufacturing to mechanised mass production providing more jobs, AI will be (potentially?) initially not providing a mass numbers of new jobs.
 
From what I'm told most of the last 10 years or so of Coke adverts have used CGI, so I'm not sure why this is different. AI is just a slightly different form of CGI.
Personally I dont have a problem with it used in films, TV or photography... for fun, or for those unable to get out shooting. They can still create images that please themselves, and possibly other.
My grip with AI is it being passed off as real. Theres thousands of YouTube videos claming to be crash at so and so, or ship docks, and they are fake AI trying to pass off as real.
 
swanseamale47 said:
From what I'm told most of the last 10 years or so of Coke adverts have used CGI, so I'm not sure why this is different. AI is just a slightly different form of CGI.
Personally I dont have a problem with it used in films, TV or photography... for fun, or for those unable to get out shooting. They can still create images that please themselves, and possibly other.
My grip with AI is it being passed off as real. Theres thousands of YouTube videos claming to be crash at so and so, or ship docks, and they are fake AI trying to pass off as real.
Click to expand...
Scott (who did the video) has discussed the effect of CGI on his industry (commercial/advertising photography), and it hasn't had the impact predicted. The main reason appears to be cost, as it's still cheaper to do it "properly" than it is to use CGI. So CGI is only used when there isn't an alternative.
 
sirch said:
I can remember people being dismissive about mobile phones and the world wide web but here we are. All of these technologies are disruptive, revolutionary and here to stay. As someone said "technology is the word we have for things that don't work ... yet"
Click to expand...
I don't...I remember mobile phones being ultra desirable before they were widespread, the issue was the cost. With the internet, those being dismissive weren't against the idea but supported the original BBS which they thought would be the future. With genAI/LLMs it's the exact opposite, these companies are having to force these tools down our throats because most people aren't bothered about them - just look at any modern W11 install and see how hard it is to get shot of Copilot.

There's also many, many fundamental problems with genAI (I hate how AI is being used for these systems which are almost the entire opposite of AI) which are not getting resolved. On a photography forum the most obvious issue is that these genAI tools function by stealing people's contents to train on, the companies have openly admitted they cannot function if they had to pay for the content. Those people who create the content cannot keep doing so for nothing if AI companies are stealing it all and already many sites have noticed drastic drops in click throughs from search engines because AI tools are stealing the content and showing it on the search results screen. Furthermore, these AI companies want to sell that stolen content back to users which is also a problem...OpenAI boast of their huge userbase yet are burning massive amounts of cash every year.

These genAI systems are also massively inefficient and they're going to get a lot worse if you look at OpenAI's 1.4 trillion dollar spending plans on new hardware. The amount of power and water being wasted is absolutely horrific and these GPUs only last a couple of years before newer models supercede them.

They're also doing massive damage to the internet with the dead internet theory feeling very accurate, I find most search engines are getting to the point they're entirely useless because there's so much AI garbage in the search results. Improvements in the AI models make them look more convincing but you quickly realise are just garbage. Unfortunately I've found an increasing number of people who think they can 'help' on forum by using a genAI tool to provide an answer to a question on a topic not realising it's worthless junk and doesn't help at all.

And now we're finding out it gets better yet because these genAI tools are now bringing about the destruction of the consumer computing market. Ram prices on PC have jumped hugely as memory companies are focusing on manufacturing memory for AI hardware instead of consumers, Crucial/Micron are pulling out of the consumer market entirely and this is going to affect everyone, Samsung won't guarantee memory to their own mobile phone division because they want to see how much they can get from AI companies. The single board Raspberry Pi's have already started putting prices up because RAM is going to be a lot more expensive for them. SSDs are already climbing in price and likely to follow ram pricing and it even affects non-ram components, motherboard sales are crashing because people are understandably not building PCs when the cost has suddenly shot up so much.

I think the fundamental problem with these genAI systems is people don't realise how stupid they are and the companies have convinced people there's a breakthrough round the corner that's going to change it all. However, these tools are little more than fancy auto predict systems hence getting it wrong is an inherent part of their process. One of the most illustrating examples of how stupid these tools are was when ChatGPT was pitched against chess....on the Atari 2600 with its beefy 1MHz CPU and 128bytes of ram. The Atari slaughtered ChatGPT demonstrating ChatGPT's lack of computational capability.

I'll give you a more accurate comparison, rather than comparing genAI to known successful technologies which bear no similarities I'll give you something much closer - cryptocurrency. Like genAI, cryptocurrency was to be the future and naysayers were reminded how many technologies in the past people had snubbed and were now dominant. Also like genAI, cryptocurrency is massively inefficient and destroyed the consumer GPU market and now where is it? People will point to the value of bitcoin as its success but it's nothing more than a gambling tool now, it serves no useful function. Remember NFTs? Also hugely hyped as the future even on photography forums and now the technology is worthless. Cryptocurrency does have some small niche uses (apart from crime which it's primarily used for) like genAI but it cannot do what it was originally promised.

Molly White who mostly covers the world of cryptocurrency summed up genAI tools extremely well:

When I boil it down, I find my feelings about AI are actually pretty similar to my feelings about blockchains: they do a poor job of much of what people try to do with them, they can't do the things their creators claim they one day might, and many of the things they are well suited to do may not be altogether that beneficial. And while I do think that AI tools are more broadly useful than blockchains, they also come with similarly monstrous costs.
Click to expand...

I'm not going to go into the potential damage these genAI tools are doing to human minds either which from early studies look quite horrific.
 
Last edited:
CGI needs a human to create the models for the idea/work so is still a creative process and employs people, AI just needs to be given a few words and is removing employment from people.
 
Gav. said:
CGI needs a human to create the models for the idea/work so is still a creative process and employs people, AI just needs to be given a few words and is removing employment from people.
Click to expand...
I think you may have missed a key point of the video, but then I'm not sure if any of the comments so far relate to the video :)
 
myotis said:
I think you may have missed a key point of the video, but then I'm not sure if any of the comments so far relate to the video :)
Click to expand...
Sorry, I was just referring to the use of CGI in ads vs AI, as it was mentioned above.
 
Gav. said:
Sorry, I was just referring to the use of CGI in ads vs AI, as it was mentioned above.
Click to expand...
Fair enough, but one of the things discussed in the video, and also discussed in his other videos is how he has had one of the busiest years ever, and a lot of it is because clients are finding AI generated ads are more expensive and not as good as those generated traditionally, except for exceptional use cases.

This mirrors what happened in his industry when CGI came out, when, like Generative ai, it was forecast as killing his work, and it didn't happen. The main explanations, he suggests, and as I posted earlier, is cost, but being able to meet clients subtle needs is another, as well as a rejection by the public of AI generated ads.

For me, this was the most interesting part of the video. But he has been mentioning it every so often for a few years now.
 
I suppose the slight framing errors in the jump cuts are done to prove he is real. :)


Actually I quite like most of what he says in his You Tube videos and smile each time he predicts a future of doom, remembering how accidentally I got out of the photographic supply business just before digital ripped the heart out of materials and chemistry sales.
 
zx9 said:
I suppose the slight framing errors in the jump cuts are done to prove he is real. :)


Actually I quite like most of what he says in his You Tube videos and smile each time he predicts a future of doom, remembering how accidentally I got out of the photographic supply business just before digital ripped the heart out of materials and chemistry sales.
Click to expand...
I like his videos but probably because I nearly always agree with what he is saying. However, I didn't think he was predicting a future of doom, but emphasising the need to change with the times.

I think he explains the jump cuts as an issue with using AI to do the editing, which he uses for his Youtube videos to maximise his income from them. I note he mentions not being able to get a sponsor for them this time, because of the impossible posting schedule the potential sponsors wanted.
 
myotis said:
I think you may have missed a key point of the video, but then I'm not sure if any of the comments so far relate to the video :)
Click to expand...
What would you say the key point is in the video then? I have watched it but he meanders all over the place and makes quite noticeable mistakes confusing completely different AI technologies so I wasn't clear on the point he was making apart from he doesn't like AI..or some types of AI. I'm also not convinced about his arguments either, he cites the dislike ratio on the Coca-Cola advert as proof of the growing backlash against genAI but are there any sales figures to show Coca-Cola have lost a lot of sales? I've never seen any and doubt it. There's also been a backlash against AI in gaming as well and one of the biggest games receiving this is Arc Raiders with Eurogamer giving it two stars out of five for its use of LLM technology for voice lines....but it's also been one of the biggest hits of the year with surprisingly large player counts so that backlash didn't go far.

Certainly I think there is truth to the claims about LLM generated artwork and video needing so much corrective work there's no point in using the genAI tool in the first place. I'd read an article on a company that had been told they needed to work with a genAI prompter instead of a digital artist to make the scene they wanted and what they found was that while the genAI tool could produce a decent image close to what they're wanting, they couldn't refine it in anyway - the first image had too few people so they asked for more but then they didn't just get more people it also changed the scene and where the people were. So they asked for slightly less people and moved to a different position but the tools can't do that, they can only keep generating scenes until one suits what's needed which may not happen whereas a digital artist could modify the scene to suit exactly what was wanted.

Although LLMs are meant to be at their strongest when coding that seems to be showing cracks with studies showing that generated code doesn't save any time because it needs more time to be corrected and check it works correctly.

Certainly I'd like to think there's some truth in what he's saying but I'm not convinced since I'm frequently surprised how people are either indifferent to LLM generated items or will keep accepting and using them even though they keep providing wrong answers. Even here there's been a number of posts defending the technology even though photographers are an area where the tools are most damaging.
 
JohnMcL7 said:
What would you say the key point is in the video then? I have watched it but he meanders all over the place and makes quite noticeable mistakes confusing completely different AI technologies so I wasn't clear on the point he was making apart from he doesn't like AI..or some types of AI. I'm also not convinced about his arguments either, he cites the dislike ratio on the Coca-Cola advert as proof of the growing backlash against genAI but are there any sales figures to show Coca-Cola have lost a lot of sales? I've never seen any and doubt it. There's also been a backlash against AI in gaming as well and one of the biggest games receiving this is Arc Raiders with Eurogamer giving it two stars out of five for its use of LLM technology for voice lines....but it's also been one of the biggest hits of the year with surprisingly large player counts so that backlash didn't go far.

Certainly I think there is truth to the claims about LLM generated artwork and video needing so much corrective work there's no point in using the genAI tool in the first place. I'd read an article on a company that had been told they needed to work with a genAI prompter instead of a digital artist to make the scene they wanted and what they found was that while the genAI tool could produce a decent image close to what they're wanting, they couldn't refine it in anyway - the first image had too few people so they asked for more but then they didn't just get more people it also changed the scene and where the people were. So they asked for slightly less people and moved to a different position but the tools can't do that, they can only keep generating scenes until one suits what's needed which may not happen whereas a digital artist could modify the scene to suit exactly what was wanted.

Although LLMs are meant to be at their strongest when coding that seems to be showing cracks with studies showing that generated code doesn't save any time because it needs more time to be corrected and check it works correctly.

Certainly I'd like to think there's some truth in what he's saying but I'm not convinced since I'm frequently surprised how people are either indifferent to LLM generated items or will keep accepting and using them even though they keep providing wrong answers. Even here there's been a number of posts defending the technology even though photographers are an area where the tools are most damaging.
Click to expand...
Mmmm, I posted a reply last night and it seems to have vanished. And yes, these videos are always a bit of a ramble.

The main point, I think, is that in spite of fears that Generative AI would have a serious detrimental impact on his business, it isn't.

He says that this has been the best year ever for business, and suggests at least part of the explanation, based on what his clients say, is that they are avoiding AI because of costs, technical limitations and the reaction from their customers to the use of AI in their adverts.

The Cocoa Cola ad, plus the comments, were used to support his own experience, and to illustrate some of the problems with AI.

At the time, CGI was seen as posing a similar threat, but in practice it didn't affect his business either, hence the comparison with CGI.

Overall, I think he is making the same points that you make in your post, but through his eyes as a high end commercial/advertising photographer.

I don't think he dislikes AI, he uses it himself for things he feels it's useful for e.g he edits his YouTube videos with AI, but not his "commercial" work. I don't dislike AI, but feel it needs to be used for the right task and carefully managed.

I especially agree with your penultimate point and made a similar, more general point in an earlier post. Despite the evidence, marketing always seems to win with these things. As a non photographic example, I saw the results of a survey somewhere (I can't remember where) that 40% of people trust the advice given by ChatGPT over the advice given by their Doctor.

I fear, that many people just don't care, and we end up with what the moneymakers, marketers, want us to have, regardless of what we need or should be wanting.

EDIT: as an afterthought, I should say that I watch a lot of Scott's videos, and he often discusses AI, so I am unavoidably looking at this video with a good idea of his experiences and opinions. This is bound to affect how I am interpreting what he is saying.

EDIT: on the same theme of the public's response to AI, I posted this earlier in the year:

 
Last edited:
sirch said:
I can still see it
Click to expand...
That's strange. I still can't see it. I can see the post I made this morning, and obviously your reply that I am replying to, but not my original reply.

I've closed and reopened the forum and shut down and re-opened Chrome, but still can't see it.

EDIT: I have looked at the post numbers, and they are all consecutive. John's post is 19, and my reply to him this morning is 20, so for me my post from yesterday doesn't seem to exist.

EDIT 2: I've now logged out and back in, but still no post

EDIT 3: fresh log in using a different computer (needed new authorisation code). but still no post
 
Last edited:
myotis said:
That's strange. I still can't see it. I can see the post I made this morning, and obviously your reply that I am replying to, but not my original reply.

I've closed and reopened the forum and shut down and re-opened Chrome, but still can't see it.

EDIT: I have looked at the post numbers, and they are all consecutive. John's post is 19, and my reply to him this morning is 20, so for me my post from yesterday doesn't seem to exist.

EDIT 2: I've now logged out and back in, but still no post

EDIT 3: fresh log in using a different computer (needed new authorisation code). but still no post
Click to expand...
I can only see your post from this morning, nothing dated from last night.
 
Gav. said:
I can only see your post from this morning, nothing dated from last night.
Click to expand...
Thanks, my original post must be in a special moderator-only part of the forum :)
 
Last edited:
You must log in or register to reply here.
Back
Top