It depends on how good they are. I'd pay a little to not have to prompt and wait for a bunch of icons, a little more to not have to curate and reroll, and more to not have to train a lora, all assuming those are buying me quality thresholds I care about.
Personally I wouldn't use ai generated images in production. To me AI generated images are simply a curiosity or toy. Maybe placeholders while you get actual art created, but to use them in a final product is just anti-humanist
I see this sort of argument a lot, but I think it's too oversimplified, there's definitely a category difference between the kinds of automation we were doing before the recent generative AI boom and that we are now doing with generative AI. The type of stuff we were automating before was largely (though obviously, not entirely) stuff that I think nobody really wanted to have to do in the first place, like manually inputting a bunch of data or what have you. If there wasn't a category difference, then this wouldn't have been a big deal worth so much VC funds and market cap.
That said, that's not my real reason to not use generative AI. My real reason to not use generative AI is that it still kinda sucks at fine details, and that annoys me greatly. These images have a great consistent style to them, but you can see that they're not really that clean. It's possible to tell at a glance, but really possible to see when you zoom in on them, especially depending on the icon. Whether this matters to you is up to taste; Personally, I'd rather have less detailed vector icons that are less technical but are very clean. If I really could do it, though, what I'd actually prefer is hand-crafted icons that are similar to these but with careful attention to detail and no weird artifacts when you pay too close of attention.
I can see that people largely don't care. Some people just have no taste and will jam ugly image generations with obvious, blatant artifacts into their blog posts; you do you. Others will use generative AI carefully in a way where you're not immediately sure if it's gen AI or not, but you probably suspect it; I kinda dislike this, but I can 100% understand it. Thiings is kind of in that group.
Very possible that some day soon genAI will be able to just produce perfect looking icons like this, no text errors, no weird artifacts, maybe even produce them in flawless looking vector SVGs or something. Maybe. For now though, it's tempting, but probably not for me.
I would happily use these as inspiration or reference, though. The broad strokes are good, it's the details that bug me to no end.
It only doesn't require work if you want a one-off amusement. If you want something specific that matches the surrounding graphics and shows exactly what you want, you're going to spend time and effort speccing it and iterating.
If you're making an icon of Tower Bridge, you're going to tag it with "London" and "bridge", so it's going to turn up in all searches for London bridge.
At this point though, the two bridges should just swap names.
I've contributed a bit to Noun Project, it's lovely. Submitted icons need to pass a manual quality gate, which does take a while, but it means that you don't run into bad AI slop or broken SVG paths as often as you do from other stock graphics services.
Capacitor looks weird. When I type "lead" is offers me to add it, but when I type "asp" it shows me "wasp" without option to add "asp" (a fish). Diode is shorted. I really don't understand the purpose of this.
So many of them look weird. The aesthetic seems to be soft, clay-like icons but then there will be lots of weirdly detailed or sharp parts.
And there's so many weird specific design choices that a human artist would not make. Why is the subway tunnel curved on one side and squared on the other side? Why does the nature journal have two different bookmarks (one of which awkwardly covers the E in NATURE)? Why would a wall outlet have one plug with a ground prong and one without? Why does the Golden Gate Bridge look like an M.C. Escher piece? Why does the bingo ball look more like a pool ball?
Also a lot of items that are very clearly a specific brand, even though the description is generic. The "smart thermostat" is a Nest. The "soccer shoe" is made by Adidas. The "smart speaker" is an Amazon Echo Dot. The "wireless earbuds" are AirPods (and for some reason there's three of them).
And then there's the blatant AI goofs; the logo and text being distorted on the HP 11c calculator, the VCR having an EPICT button and a REE jack, the TARDIS reading "POLIC BOX", the egg timer reading "30 10 10 10 15", the playing cards having two aces of clubs (one of which is red).
I like this handmade/clay aesthetic, but it completely falls apart when it's obvious a human hasn't touched it. If I want handmade icons, I'll pay an artist for them; if I want AI slop, I can generate it myself.
For the laundromat it couldn't decide if it's the laundromat building or the washing machine itself. Reminds me of Rollercoaster Tycoon graphics but unintentional and inconsistent.
I am interested in guidance for generating a custom icon like this with a similar level of style consistency. I know there are some resources out there with guidance, but does anyone know any really good ones?
This is absolutely not my area of expertise, and I can't fully vouch for how well it works, but I was looking for a similar solution for consistency across website graphics and came across this walkthrough for consistent game assets:
Generate something you like the style of, then run it back through a vision model and ask it to describe the image without describing the subject, append that to a batch of icon descriptions, easy custom icons. (figure out transparency at this step) Then ask it for an imagemagick script that will convert them into needed sizes and formats.
very manual yes, but gives a lot of control
[edit] the images seem to be made with a transformers model rather than diffusion (e.g. DallE3 vs MidJourney) which is mostly proprietary for now.
Spinosaurus is the inaccurate one from Jurassic Park 3 and Tyrannosaurus has inaccurate hand rotation. Velociraptor and Dilophosaurus are also from Jurassic Park and highly inaccurate. For shame!
This page feels like one of those times that touch and drag inertia would really be a benefit. It feels so clumsy to navigate on my phone but looks so close to being a delight.
Well… I guess if I’m limited to only discussing the most significant part of any issue, it would have to be that none of this matters, agency is an illusion, and we should all sell our computers and find people to feed.
But why does it exist? Will it be around in 6 months? What is the backstory / creator of this project?
It's odd that they use their own license language, and doesn't use something explicit like CC-BY-NC. It appears to be a non-commercial, non-distribution license, but otherwise free for personal and commercial use...?
Appears that the site is funded through being able to purchase sponsorship spots for $10 / mo.
Confusingly, the about page suggests that this is an AI-driven storytelling app: https://www.thiings.co/about. Maybe that's a vestige of prior iteration of the app, though.
The concept minds me of the million dollar homepage, where you could pay a dollar per pixel of advertisement. Here, you're paying a dollar (that's a guess; i didn't check) to get a new object listed in the potentially infinite grid, which gives you a unique URL for that object, an emoji-like reference image, and a one-paragraph description for aliens, were they to land on earth and ask what that object was for.
Basically, looks like an art project that monetizes participation.
Love this beautiful pattern wallpaper generator but I have two questions: How do I save a pattern (rightclick download doesn't work) and can you make the background color adjustable? Thanks!
This is a great idea, but the monetization model will not stand the pressure of market forces. Icons is a race to the bottom, even with relatively high-cost human labor
I think this is best seen as an art project, not as a business idea. I'm also assuming that the cost to run such an app are near zero, so the break-even cost would probably be, like, one person a month generating one new thing.
I feel like this has been regenerated since this comment was posted - there's no scissors here now, but there _is_ an impossible thread skein (or whatever you call it for thread at least).
Oh, wow, they must have regenerated it. When I was there, it had some busted scissors with metal bars sticking out in both directions around the fingerholes.
Kind of sloppy, though. An impossible Rubik's cube, a sudoku puzzle with incorrect numbers, a mirror showing a non-reversed reflection, a domed White House, a US flag with 32 stars...
> Thiings is a growing collection of 1,900+ free AI-generated 3D icons
"Free", but downloading the entire collection requires a payment.
And according to the terms:
> 4. Content
> You retain all rights to your content. By using our service, you grant us a license to host and display your content.
But then they advertise that anyone can download and use these pictures? Under what license?
If they're AI-generated, just grab them and ignore the terms. What are they gonna do, claim copyright?
I see someone has heard from their legal department on AI tools.
Seriously, why would I pay for someone else to generate AI images?
It depends on how good they are. I'd pay a little to not have to prompt and wait for a bunch of icons, a little more to not have to curate and reroll, and more to not have to train a lora, all assuming those are buying me quality thresholds I care about.
Personally I wouldn't use ai generated images in production. To me AI generated images are simply a curiosity or toy. Maybe placeholders while you get actual art created, but to use them in a final product is just anti-humanist
I've been automating tasks away from humans for going on 25 years now. Unadulterated misanthropy.
I see this sort of argument a lot, but I think it's too oversimplified, there's definitely a category difference between the kinds of automation we were doing before the recent generative AI boom and that we are now doing with generative AI. The type of stuff we were automating before was largely (though obviously, not entirely) stuff that I think nobody really wanted to have to do in the first place, like manually inputting a bunch of data or what have you. If there wasn't a category difference, then this wouldn't have been a big deal worth so much VC funds and market cap.
That said, that's not my real reason to not use generative AI. My real reason to not use generative AI is that it still kinda sucks at fine details, and that annoys me greatly. These images have a great consistent style to them, but you can see that they're not really that clean. It's possible to tell at a glance, but really possible to see when you zoom in on them, especially depending on the icon. Whether this matters to you is up to taste; Personally, I'd rather have less detailed vector icons that are less technical but are very clean. If I really could do it, though, what I'd actually prefer is hand-crafted icons that are similar to these but with careful attention to detail and no weird artifacts when you pay too close of attention.
I can see that people largely don't care. Some people just have no taste and will jam ugly image generations with obvious, blatant artifacts into their blog posts; you do you. Others will use generative AI carefully in a way where you're not immediately sure if it's gen AI or not, but you probably suspect it; I kinda dislike this, but I can 100% understand it. Thiings is kind of in that group.
Very possible that some day soon genAI will be able to just produce perfect looking icons like this, no text errors, no weird artifacts, maybe even produce them in flawless looking vector SVGs or something. Maybe. For now though, it's tempting, but probably not for me.
I would happily use these as inspiration or reference, though. The broad strokes are good, it's the details that bug me to no end.
ChatGPT 4o image generation is pretty good at icons/ui/ux, likewise with text, might require a re-roll and not one shot long texts but still...
also, there's some genAI for SVG's too, so you could -modify- or tune-up these after a first AI gen if thats' your choice.
imho these are tools and will be useful to augment current workflows, a human will be always required to -edit/curate- because thats' what we do
I use them in some products that don't exist otherwise, or wouldn't have any images
I’ve often found that no images would be preferable to AI-generated ones.
For sure, there are many such cases. But I get a lot of positive feedback where I've used them.
Because it requires work and you place value on your time
Does it really require work though?
Could you share a similar library you've built without any work?
Do most people need 20k images, or just a handful?
No but I can point to hundreds on HN.
Please provide a couple, I am genuinely interested.
Uh, yes?
What work?
QA and organization of the assets...
The entire point of AI is that it doesn't require work, you just type a prompt in and it does the work for you
Prompting is still work; it requires thought and oftentimes iteration.
It requires time, not work
It only doesn't require work if you want a one-off amusement. If you want something specific that matches the surrounding graphics and shows exactly what you want, you're going to spend time and effort speccing it and iterating.
Along the same lines, less "cute", but far more extensive - https://thenounproject.com
Noun Project is fantastic. It's been around for at least a decade and it's hard to find things with no icons these days.
I notice it suffers from the same London Bridge problem, do people never learn? https://en.wikipedia.org/wiki/London_Bridge_(Lake_Havasu_Cit...
If you're making an icon of Tower Bridge, you're going to tag it with "London" and "bridge", so it's going to turn up in all searches for London bridge.
At this point though, the two bridges should just swap names.
I've contributed a bit to Noun Project, it's lovely. Submitted icons need to pass a manual quality gate, which does take a while, but it means that you don't run into bad AI slop or broken SVG paths as often as you do from other stock graphics services.
Capacitor looks weird. When I type "lead" is offers me to add it, but when I type "asp" it shows me "wasp" without option to add "asp" (a fish). Diode is shorted. I really don't understand the purpose of this.
So many of them look weird. The aesthetic seems to be soft, clay-like icons but then there will be lots of weirdly detailed or sharp parts.
And there's so many weird specific design choices that a human artist would not make. Why is the subway tunnel curved on one side and squared on the other side? Why does the nature journal have two different bookmarks (one of which awkwardly covers the E in NATURE)? Why would a wall outlet have one plug with a ground prong and one without? Why does the Golden Gate Bridge look like an M.C. Escher piece? Why does the bingo ball look more like a pool ball?
Also a lot of items that are very clearly a specific brand, even though the description is generic. The "smart thermostat" is a Nest. The "soccer shoe" is made by Adidas. The "smart speaker" is an Amazon Echo Dot. The "wireless earbuds" are AirPods (and for some reason there's three of them).
And then there's the blatant AI goofs; the logo and text being distorted on the HP 11c calculator, the VCR having an EPICT button and a REE jack, the TARDIS reading "POLIC BOX", the egg timer reading "30 10 10 10 15", the playing cards having two aces of clubs (one of which is red).
I like this handmade/clay aesthetic, but it completely falls apart when it's obvious a human hasn't touched it. If I want handmade icons, I'll pay an artist for them; if I want AI slop, I can generate it myself.
For the laundromat it couldn't decide if it's the laundromat building or the washing machine itself. Reminds me of Rollercoaster Tycoon graphics but unintentional and inconsistent.
Then don't use it. Simple as.
I second that.
[dead]
The entry for London Bridge is wrong, Tower Bridge is depicted.
Staff is also wrong: https://www.thiings.co/things/staff - shows a group of people but describes a stick.
And some of those people have mouths and noses, some don't. Creepy image.
The needle is all wrong, the thread doesn't go through the eyelet.
https://www.thiings.co/things/needle-unott0
Take a look at the scissors on the cross stitch kit: https://www.thiings.co/things/cross-stitch-kit
The puzzle cube is colored wrongly
https://www.thiings.co/things/puzzle-cube
Only half of all wall outlets need the ground prong: https://www.thiings.co/things/wall-outlet
And this decidedly is not a square not: https://www.thiings.co/things/square-knot
Some wall outlets in some regions do combine a two-prong and three-prong socket.
Though I don't know if that applies to the region that has the D: face sockets.
The image for Physics shows an impossible Newton's Cradle. The balls should be, well, cradled between two strings.
https://www.thiings.co/things/physics
It instead goes through the needle itself... Just AI things
Golden Gate bridge is apparently designed by MC Escher: https://www.thiings.co/things/golden-gate-bridge
I am interested in guidance for generating a custom icon like this with a similar level of style consistency. I know there are some resources out there with guidance, but does anyone know any really good ones?
This is absolutely not my area of expertise, and I can't fully vouch for how well it works, but I was looking for a similar solution for consistency across website graphics and came across this walkthrough for consistent game assets:
https://runware.ai/blog/creating-consistent-gaming-assets-wi...
Maybe helpful?
Generate something you like the style of, then run it back through a vision model and ask it to describe the image without describing the subject, append that to a batch of icon descriptions, easy custom icons. (figure out transparency at this step) Then ask it for an imagemagick script that will convert them into needed sizes and formats.
very manual yes, but gives a lot of control
[edit] the images seem to be made with a transformers model rather than diffusion (e.g. DallE3 vs MidJourney) which is mostly proprietary for now.
Me as well. The one thing I've struggled with is consistently getting a combination of object-only and transparent background.
one good technique is to pass the style guide as a json, where you define materials, lightning, perspective, etc
you can even use a vision model to generate the style guide for you
Very clever thank you
Of course it has “Reaper Drone” as well as several different tanks and war planes, but not, for instance, “dildo”.
Somewhat relevant GitHub project, for generating Genmoji-style images: https://github.com/EvanZhouDev/open-genmoji
How did you get fireship to narrate?
(more seriously, the timing on the playback does not seem to account for load time, audio may start a couple words in if not instantly loaded)
Spinosaurus is the inaccurate one from Jurassic Park 3 and Tyrannosaurus has inaccurate hand rotation. Velociraptor and Dilophosaurus are also from Jurassic Park and highly inaccurate. For shame!
This page feels like one of those times that touch and drag inertia would really be a benefit. It feels so clumsy to navigate on my phone but looks so close to being a delight.
I also want to pinch zoom out to see the full database, but I guess they won't give it out for free.
Really, that's your takeaway from all of this?
Well… I guess if I’m limited to only discussing the most significant part of any issue, it would have to be that none of this matters, agency is an illusion, and we should all sell our computers and find people to feed.
But touch drag inertia is a close second.
Cool site!
But why does it exist? Will it be around in 6 months? What is the backstory / creator of this project?
It's odd that they use their own license language, and doesn't use something explicit like CC-BY-NC. It appears to be a non-commercial, non-distribution license, but otherwise free for personal and commercial use...?
Appears that the site is funded through being able to purchase sponsorship spots for $10 / mo.
Confusingly, the about page suggests that this is an AI-driven storytelling app: https://www.thiings.co/about. Maybe that's a vestige of prior iteration of the app, though.
The concept minds me of the million dollar homepage, where you could pay a dollar per pixel of advertisement. Here, you're paying a dollar (that's a guess; i didn't check) to get a new object listed in the potentially infinite grid, which gives you a unique URL for that object, an emoji-like reference image, and a one-paragraph description for aliens, were they to land on earth and ask what that object was for.
Basically, looks like an art project that monetizes participation.
Love this beautiful pattern wallpaper generator but I have two questions: How do I save a pattern (rightclick download doesn't work) and can you make the background color adjustable? Thanks!
This is a great idea, but the monetization model will not stand the pressure of market forces. Icons is a race to the bottom, even with relatively high-cost human labor
I think this is best seen as an art project, not as a business idea. I'm also assuming that the cost to run such an app are near zero, so the break-even cost would probably be, like, one person a month generating one new thing.
good point on break-even cost.
I can't consider this an art project when the creator has integrated:
- A lifetime membership fee for downloading all icons
- A sponsorship subscription option that lets subscribers integrate their own icon (ad) into the collection
- A credits system for generating new icons using AI
Nothing against creators getting paid, just saying that this is monetized to the gills and looks more like an indie hackers thing than an art project.
Super helpful, i was just looking for some 3d elements of bonfires and dolphins to go with my apps aesthetic. the timing could not have been better :D
thanks!
They're AI generated and look quite nice, no obvious artifacts (messed up text). Not sure what the use case is though.
The scissors here are not what I'd describe as "no obvious artifacts": https://www.thiings.co/things/cross-stitch-kit
> https://www.thiings.co/things/cross-stitch-kit
I feel like this has been regenerated since this comment was posted - there's no scissors here now, but there _is_ an impossible thread skein (or whatever you call it for thread at least).
I don't see scissors on this page, although the yarn coloring is unlikely.
Oh, wow, they must have regenerated it. When I was there, it had some busted scissors with metal bars sticking out in both directions around the fingerholes.
Kind of sloppy, though. An impossible Rubik's cube, a sudoku puzzle with incorrect numbers, a mirror showing a non-reversed reflection, a domed White House, a US flag with 32 stars...
I treat them kind of like icons/emoji, which often don't have a lot of fidelity due to size constraints.
I've been downloading them to my iPhone and converting a number of them to stickers, which can be used identically to emoji in iMessage.
It was weird reading the description for gigan as a cyclops but the image showing 2 eyes.
So you can pretend you are an alien learning about human culture.
The hammock one is non-euclidean.
Searching for "baby carrier" is hilarious
Love the website design, but that rsc lag is painful.
I searched for "hot dog" and got "Oxford University".
There's a Five Guys near the University of Oxford, so that seems an appropriate search return to me.
so good
So cool!