Live data from Hacker News

AI-Generated Images Discourage Me from Reading Your Blog

nelson.cloud

461–470 of 493 posts

Re: AI-Generated Images Discourage Me from Reading Your Blog

#461

Earlier quoted context omitted.

Right now your actual argument is just really your opinion. You can avoid actually examining it closely by claiming logical fallacies, but again we're just talking about your opinion, you're not making a logical argument.

The colloquial definition of "creative" and "art" are not solely an opinion, it's a tangible thing you can explore and investigate if you're motivated to do so. The inability of LLMs to express creativity is also not an opinion. The differences between film directors and LLM prompters are also real things that we could discuss. But a meaningful discussion requires good faith, not dismissal to disingenuous argument.

You're acting as though the ordinary meaning of the words settle the controversy, but that's exactly what's contested here.

Whatever definition you bring into scope will not resolve it because a colloquial definition only describes how a word is commonly used, and doesn't establish rule for every disputed case, which is what we have here.

Definitions are generally not normative, but even if we agreed on one, it would not exclude a works solely because of the tool used to produce it. But you are saying that every work produced using AI is illegitimate as art, regardless of its intention, composition, context, expressive content, or any other quality ordinarily associated with art.

What necessary property does accepted art possess that no AI-assisted work could possibly possess? Or to state it another way closer to my original question (which was not disengenuous), what additional human contribution makes photography or CGI art, and why couldn’t an AI-assisted process contain that contribution?

Until that property is identified, saying AI images cannot be art is largely just a stated conclusion.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#462

Earlier quoted context omitted.

> What did they Photoshop? Probably some creative work that they or somebody else made.... If all you did was write a prompt for an image Most serious AI artist are not just writing a prompt for an image. They usually have their own process and can describe it in the same detail any trad artist would. But to your point about Photoshop/autotune, the basis of most AI artwork pipelines is some other creative work, like…

I use AI all the time and do some paid work as an artist. Although I don't use AI directly in my own work, I am very familiar with the workflows in 2D, 3D, and video. > you didn't create anything in this case. Not the images at least. Which then casts doubt on the rest of the writing in said blog post. The topic is "AI-generated images" in blog posts. Which, no, most of those aren't being generated with any of the mo…

> The topic is "AI-generated images" in blog posts. Which, no, most of those aren't being generated with any of the more sophisticated techniques you're listing. And if they are, they're less likely to be the images people are complaining about and that are being discussed here.

The person you originally replied to scoped their comment to "decent usage", which to me implied some more process than just prompting.

> Does "synth" mean synthesized music?

Yes.

> A synthesizer is generally deterministic and the musician is making a decision about exactly what nodes are being played.

Generally but not always. There's plenty of synthetic music that leverage stochastic components. The stochasticity is a useful design constraint and is part of the artform. There's nothing that says art has to be deterministic or that the artist can't leverage nondeterminism in their artwork.

> Text prompt-based generation is a much higher level of abstraction which often makes several compositional and stylistic decisions for the user, even with advanced prompting.

There's also a whole genre of synthetic music called "Live Coding" which is typing programs to create synthetic music in front of an audience. There's also nothing that says art is only about making low level decisions.

> Go ahead and use AI-images in your blog post, I'm not gatekeeping or going to stop you. If they're so great, use a bunch of them everywhere. But, I (and people in this thread or that I know) will assume it's blogspam and click away.

It all depends on who you're looking to attract. The people I've met in AI focused communities have been awesome, always looking to experiment, try new things. The people against them I find very conservative and resistant to change (not saying this is you, or that it's necessarily bad). Which is fine if that's your audience, you have to know your readership. It's sort of like the difference between startup culture and enterprise culture. One is known for being nimble and innovative, the other for being reliable and capable. Positive signals to one are anti signals to the other. Threats like "I and people I know and the people in this thread won't frequent your blog" are met with shrugs, because that just means you're not the target audience.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#463
post #8

To me there are different motivations to include imagery and some of them are just fine with AI generation. If the imagery actually goes to the substance of helping to illustrate or explain the actual content then it seems very superficial to reject it based on how it was generated. In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way. The problem com…

> In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way.

That is not at all what the author wrote. They said they would rather see a shitty hand drawing than an AI image, they’re not “insisting” you make a hand drawn image. You could just as well use one of the thousands of free stock photos out there just like blogs have been doing for years before AI images.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#464
post #61

There’s also a class of “detailed agent written technical blog” that’s you can just scan and get the shape that it’s long paragraphs, tables of statistics, and dry LLM writing. You notice the blog has one or more such posts a day. It’s a turnoff immediately if someone shares what their agent wrote. It also has happened a few times they contain a fair number of misrepresentations and mistakes.

On the other hand, it could be great to have collapsed sections clearly marked as LLM with the name of the model, the prompt, the assumption the author somewhat checked it and didn't use it to write the article. You can put it inline or at the end of the article: references, citations, further reading, "AI". (It might also be the only way to stop training models on their own output)

I don't see a lot of value in that, but one could supply a lot of semi structured extra information which could aid the readers LLM to answer questions the reader might have with more accuracy / factual grounding.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#465
post #8

To me there are different motivations to include imagery and some of them are just fine with AI generation. If the imagery actually goes to the substance of helping to illustrate or explain the actual content then it seems very superficial to reject it based on how it was generated. In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way. The problem com…

> In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way.

Would you also say this about the writing? It seems to me from your description that you consider drawing something to be harder than writing something.

Which is fine, but to be fair, I'd also prefer that people who find drawing easier than writing, to not outsource their writing to an LLM. By which I mean I would prefer not to read that.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#466

Earlier quoted context omitted.

The most important thing to keep in mind, when considering AI images for helping to illustrate or explain the actual content, is that AI-generated images have a big trust problem. Take a look at the Google Image Search results for 'Engine Diagram' https://www.google.com/search?q=Engine+Diagram&tbm=isch In the first row I get two correct-looking diagrams, then a load of AI-generated ones. Three cylinder heads for four…

To be fair, the problems you describe existed before AI, though I concede it has gotten worse. What baffles me is why anyone would think it a good idea to post and host these images? What’s that upside?

> To be fair, the problems you describe existed before AI, though I concede it has gotten worse.

I do think it's fair to say it's gotten a lot worse.

Before AI-generated imagery, an hilariously bad (yet human created) engine diagram as described above, might actually be shared among my friends for laughs. Because it didn't happen that often. Now it happens a lot more often, and they aren't human created (which was only funny for a very short while).

Re: AI-Generated Images Discourage Me from Reading Your Blog

#467

Earlier quoted context omitted.

not everyone who can think and solve interesting problems can write prose well, but there's still a lot of complaining about AI-generated text.

One of the most tedious ideas of recent years is the "great engineers are great writers" meme, unsurprisingly spread by writers.

Unfamiliar with this meme.

But I'd still rather read a badly human engineer written blog post than an LLM generated one.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#468

Earlier quoted context omitted.

> You're setting yourself up for a lonely life. What a ridiculous conclusion jump. It's the complete opposite of that. I prefer human things, human interaction, with human beings involved, which in some way reflect the human experience.

No. This anti-AI attitude is getting old and tired. I'm getting sick of people insulting the work that I do, calling it lazy (it's not), and saying I should do things in the manner they prefer. I'm also sick of the childish death threats. If you don't want to interact with me because of AI, good riddance.

[dead]

Re: AI-Generated Images Discourage Me from Reading Your Blog

#469
post #8

To me there are different motivations to include imagery and some of them are just fine with AI generation. If the imagery actually goes to the substance of helping to illustrate or explain the actual content then it seems very superficial to reject it based on how it was generated. In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way. The problem com…

> In fact it is kind of stupid to insist someone slavishly draw something by hand that could have been done an easier way.

Its not. Putting in the effort makes sure there was a reason for it in the first place and gives the author to re-evaluate his assumptions.

Re: AI-Generated Images Discourage Me from Reading Your Blog

#470

Earlier quoted context omitted.

> Yesterday I encountered AI-generated food photos on a menu for the first time. I immediately felt grossed out before even realising why. Their generated focaccias didn't even look like food, just off-putting trypophobia triggers. Also, that's just misleading. When I see pictures of food on a menu, I expect them to be real pictures of an example of the actual dish I'd get (perhaps made a little more carefully with m…

Not that either are OK, but McDonalds basically stages their food to make it look more appealing, adding stuff that isn't in the product to make it look nicer. https://youtu.be/pSYw8LTXXJM?is=FsJwTFKUkeDt1Z2Y Tangentially, I kinda like the idea of Doordashs crowd sourced food images. Take a Pic of your food before you eat and get like a $5 discount if it is chosen. Customers are not likely to manipulate the product (…

McD's is also not considered quality food so I don't think this example goes against the association of fake pics -> no thanks. But the practice also becomes more egregious when applied to more niche foods where the customer doesn't already have an idea of what the food could look like.
Post reply on HN