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.
AI thumbnails outperform human thumbnails on YouTube by a lot, I've seen click-through rate spreads of 4%+ (~5% -> ~9% in AB tests). Yet people will do drive by comments trying to excoriate creators for using AI thumbnails. It's really quite ridiculous.
A lot of any real gain will be due to novelty, and that will wane once you've been doing it a short time (or once other people in your field are doing it so you all go back to looking the same).
I suspect a lot of the gain isn't even that useful to the channel: it being people clicking to complain. You might not see that mean actual complaints because many will see the existing ones and either just upvote them or move on without doing anything (their chance to repeat their regular comment to get a few upvotes having been taken by someone already having said it).
Yes, and this is a bad thing. It's bad that it gets rewarded and it's bad that creators whose main content is good feel compelled to do it. And I would watch all of this stuff just the same, based on the topic matter and the channel's reputation, without them doing these things, as long as YouTube had put a trial of that content in front of me in the first place. And I get no way to actually give that feedback on YouTube.
There are plenty of channels I'll watch semi-regularly that have god-awful thumbnails. But the same is true of almost everything that I immediately click "don't recommend channel" on.
I'm saying that it at best drives a temporary increase in engagement, then you look even more like everyone else than you did before because everyone is doing the same thing with the same AIs. And while that engagement counts in the immediate stats, it might result in a drop overall as some mark you as a channel they don't want to visit again. Then again, I don't think many youtubers are relying on discerning viewers so maybe they will get a boost in the hollow attention the crave.
It’s their job, their source of income (or they’re aspiring to it). Making bait thumbnails is a necessary evil. If they don’t they’re leaving money on the table for no reason.
If they AB test some thumbnails and the AI ones get more views/engagement then of course they are going to do that.
I think the overall premise of “if AI thumbnails are even getting more engagement, it must be from haters” is faulty to begin with. It just fits your personal narrative. We simply don’t know.
Peers Morgan's source of income, and hence his job, is being rage bait in (vaguely) human form. A pound shop Alex Jones. It's being his job doesn't mean I have to respect it, or him. See also: loan sharks, living statues, influencers generally.
You seemed to be reacting to my low opinion of the practices of attention dredgers, as if I'm bad for holding that opinion and a should respect all career choices and income generation attempts. Perhaps I read that quite wrong.
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.
CuriouslyC · · focus · HN ↗
dspillett · · focus · HN ↗
I suspect a lot of the gain isn't even that useful to the channel: it being people clicking to complain. You might not see that mean actual complaints because many will see the existing ones and either just upvote them or move on without doing anything (their chance to repeat their regular comment to get a few upvotes having been taken by someone already having said it).
phoghed · · focus · HN ↗
whattheheckheck · · focus · HN ↗
phoghed · · focus · HN ↗
zahlman · · focus · HN ↗
There are plenty of channels I'll watch semi-regularly that have god-awful thumbnails. But the same is true of almost everything that I immediately click "don't recommend channel" on.
dspillett · · focus · HN ↗
phoghed · · focus · HN ↗
It’s their job, their source of income (or they’re aspiring to it). Making bait thumbnails is a necessary evil. If they don’t they’re leaving money on the table for no reason.
If they AB test some thumbnails and the AI ones get more views/engagement then of course they are going to do that.
I think the overall premise of “if AI thumbnails are even getting more engagement, it must be from haters” is faulty to begin with. It just fits your personal narrative. We simply don’t know.
dspillett · · focus · HN ↗
phoghed · · focus · HN ↗
dspillett · · focus · HN ↗
> It’s their job, their source of income
You seemed to be reacting to my low opinion of the practices of attention dredgers, as if I'm bad for holding that opinion and a should respect all career choices and income generation attempts. Perhaps I read that quite wrong.
phoghed · · focus · HN ↗
I didn’t say anything like that, you pulled that feeling out of yourself.
I’m just pointing out that they are doing it for money. Most white collar workers are using AI as part of their job these days.