A lot of people in the comments say that the supposedly better examples in the article are still “horrible” because they can tell they’re AI and point to things that the AI did “wrong” and a human artist would do better.
That may be true if you’re working with a top of the line artist, but in my experience the average freelance designer online is significantly worse than AI. I’ve worked with dozens of budget friendly designers on Fiverr and consistently get way worse results than I’ve gotten from AI models. Will a great human designer produce better output than an AI? Maybe, but a small local event or business probably won’t have the budget to hire them or know where to find them. It’s cheaper, faster and easier to use AI to generate something that’s good enough for their purposes.
People also point out all the little flaws in the “good” examples and cite them as obvious tells that they were AI generated. I think it would be interesting to see if people could reliably distinguish between AI generated flyers and human made flyers. Sure, sometimes it’s obvious like when the AI adds those weird, cartoony people. But if the model is told not to create designs like that and stick to the styles used in the second half of the article, I expect people would have a much harder time distinguishing the human designs from the AI designs.
There aren't small problems in these posters though. They're huge, obvious, and an indicator that the one who made the flyer didn't care whatsoever.
When it includes an American Flag with more than 50 stars + the bottom row is cut off, you know they didn't even bother to proofread it.
This is the kind of garbage we're suffering with.
Are you aware we don’t all get the same results from a google image search? Or do you seriously believe if I search the same keywords we will be looking at the exact same images?
True. But an organizer who doesn't care, will also not be willing to pay for good artwork. AI haters should put money where their mouths are and fund human top artwork.
You're really overestimating how many people can draw something "decent" in Inkscape. Most likely they will take a template. That's not necessarily better than AI either.
I don't claim to be an artist and a template is still done with the caring eye of a professional. It's way better than an AI generated poster. Or at the very least, there's many less instances where I've seen overly used templates than I have badly designed AI generated designs.
It's not really about that, and it's not really a dunk. What I'm trying to do is contrast the reality of what is actually being produced and bought by human beings, and what the parent commenter was suggesting was not caring at all about your product or brand. It was extremely convenient that the exact critique they were leveling against this small business poster could also be leveled at such a high-profile product.
Those are, comparatively, small problems for the medium of a poster.
You have to understand that 99% of people who will see the poster will view it for <5 seconds. They wont sit there and count the stars.
And, of the people who do examine the poster closely, they’re looking at the text. They want to see the date, time, event details. The background is background.
But what is the point of the background then? I mean, it seems like we're admitting that the design doesn't matter, then why keep up the ruse? Let's drop design, and keep just plain textual information. There is no substance to it, it's just performative, a simulacrum of design.
I would think to catch people's eye so they look in the first place, sort of similar to an ad. Which is also why we see AI slop work well in ads. And before AI slop, we had corpo slop. Just bright colors and wacky animated characters who all fit that archetype where their limbs are too big. So I do think the design takes a back seat, at least for the common case. Of course there are exceptions, and those exceptions are exceptional, but for most purposes slop works fine.
This is the important part. Right now, AI is a tool. You can use it to generate high quality output, or you can use it to generate slop. If the output is slop, that is a reflection of the person using the tool, not AI.
I get AI slop from employees sometimes. If they provide poor quality work that hasn't been proofread, that is an employee problem, not an AI problem.
I wholeheartedly agree. We are heading for troubled waters.
All I can do is hold the people around me accountable for doing high quality work. But everywhere I look, there are too much incentive to just let AI do the thinking for you. My generation will be okay for the most part, since we spent decades thinking prior to AI. It is unclear if the next generations, that haven't already learned how to think, will ever learn.
Not every poster has that issue as you can see from the posting. It is a simple matter of caring enough to look at the output in more detail. These are issues that can be overcome.
Sure, but the basic value proposition of AI is that you don't have to care. These issues could be overcome, but they aren't, because the whole way the technology is promoted and sold guides the user away from caring.
One of the things that keeps circling in my mind when it comes to LLMs and similar transformer architectures is they are technically very impressive, and some of the things they can do are amazing, but there seems to be a compulsive need by the executives of corporations producing them to lie about their capabilities.
I suspect that what drives the compulsive lying and exaggeration is that they know what LLMs can actually do is not worth the development and operating costs (at least for now), so, like Bernie Madoff, they pretend that they're more capable than they are in the hopes that the next round of investment will somehow bridge the gap and let them actually produce what they've been falsely promising their investors they can do, and make the whole thing worthwhile.
ajjenkins · · focus · HN ↗
That may be true if you’re working with a top of the line artist, but in my experience the average freelance designer online is significantly worse than AI. I’ve worked with dozens of budget friendly designers on Fiverr and consistently get way worse results than I’ve gotten from AI models. Will a great human designer produce better output than an AI? Maybe, but a small local event or business probably won’t have the budget to hire them or know where to find them. It’s cheaper, faster and easier to use AI to generate something that’s good enough for their purposes.
People also point out all the little flaws in the “good” examples and cite them as obvious tells that they were AI generated. I think it would be interesting to see if people could reliably distinguish between AI generated flyers and human made flyers. Sure, sometimes it’s obvious like when the AI adds those weird, cartoony people. But if the model is told not to create designs like that and stick to the styles used in the second half of the article, I expect people would have a much harder time distinguishing the human designs from the AI designs.
VCFundedGenYer · · focus · HN ↗
When it includes an American Flag with more than 50 stars + the bottom row is cut off, you know they didn't even bother to proofread it.
This is the kind of garbage we're suffering with.
qarl · · focus · HN ↗
dgellow · · focus · HN ↗
qarl · · focus · HN ↗
The first one I clicked on is at a stock photo site that says:
Every visual at iStock is crafted through human expertise, creativity, and ingenuity. We do not allow AI-generated visuals in our library.
dgellow · · focus · HN ↗
qarl · · focus · HN ↗
Did you even look?
HappMacDonald · · focus · HN ↗
Claim: A vast majority of human-made clip art that involves flags as tertiary elements will fail to render them in a vexillologically perfect manner.
FooBarWidget · · focus · HN ↗
dgellow · · focus · HN ↗
FooBarWidget · · focus · HN ↗
dgellow · · focus · HN ↗
johnnyanmac · · focus · HN ↗
HappMacDonald · · focus · HN ↗
.. except when those are also AI generated
account42 · · focus · HN ↗
rpdillon · · focus · HN ↗
[dead]
dgellow · · focus · HN ↗
Step 2-3: sell anything to your cult followers, doesn’t matter how good or bad it is
Step 4: profit
Simple recipe for success
zahlman · · focus · HN ↗
rpdillon · · focus · HN ↗
preg_match · · focus · HN ↗
You have to understand that 99% of people who will see the poster will view it for <5 seconds. They wont sit there and count the stars.
And, of the people who do examine the poster closely, they’re looking at the text. They want to see the date, time, event details. The background is background.
dsego · · focus · HN ↗
preg_match · · focus · HN ↗
richk449 · · focus · HN ↗
I get AI slop from employees sometimes. If they provide poor quality work that hasn't been proofread, that is an employee problem, not an AI problem.
johnnyanmac · · focus · HN ↗
richk449 · · focus · HN ↗
All I can do is hold the people around me accountable for doing high quality work. But everywhere I look, there are too much incentive to just let AI do the thinking for you. My generation will be okay for the most part, since we spent decades thinking prior to AI. It is unclear if the next generations, that haven't already learned how to think, will ever learn.
doug_durham · · focus · HN ↗
wiml · · focus · HN ↗
GolfPopper · · focus · HN ↗
I suspect that what drives the compulsive lying and exaggeration is that they know what LLMs can actually do is not worth the development and operating costs (at least for now), so, like Bernie Madoff, they pretend that they're more capable than they are in the hopes that the next round of investment will somehow bridge the gap and let them actually produce what they've been falsely promising their investors they can do, and make the whole thing worthwhile.
kQq9oHeAz6wLLS · · focus · HN ↗
eMpHaSe · · focus · HN ↗
zahlman · · focus · HN ↗
Remember when this sort of ascending tricolon was supposed to be a dead giveaway for LLM-written prose?