
I just want to say this isn't just amazing -- it's my new favorite map of NYC.
It's genuinely astonishing how much clearer this is than a traditional satellite map -- how it has just the right amount of complexity. I'm looking at areas I've spent a lot of time in, and getting an even better conceptual understanding of the physical layout than I've ever been able to get from satellite (technically airplane) images. This hits the perfect "sweet spot" of detail with clear "cartoon" coloring.
I see a lot of criticism here that this isn't "pixel art", so maybe there's some better term to use. I don't know what to call this precise style -- it's almost pixel art without the pixels? -- but I love it. Serious congratulations.
Author here, and to reiterate another reply - all of the critique of "pixel art" is completely fair. Aesthetically and philosophically, what AI does for "pixel art" is very off. And once you see the AI you can't really unsee it.
But I didn't want to call it a "SimCity" map, though that's really the vibe/inspiration I wanted to capture, because that implies a lot of other things, so I used the term "pixel art" even though I figured it might get a lot of (valid) pushback...
In general, labels and genres are really hard - "techno" to a deep head in Berlin is very different than "techno" to my cousins. This issue has always been fraught, because context and meaning and technique are all tangled up in these labels which are so important to some but so easily ignored by others. And those questions are even harder in the age of AI where the machine just gobbles up everything.
But regardless, it was a fun project! And to me at least it's better to just make cool ambitious things in good faith and understand that art by definition is meaningful and therefore makes people feel things from love to anger to disgust to fascination.
just FYI, there are some very obvious inaccuracies (stitching artefacts?) in this map, especially noticeable e.g. on Roosevelt Island around Roosevelt Bridge, or naround Pier 17 next to Brooklyn Bridge
There's a strange error where half the Broadway Junction subway station just... disappears. And seems to be replaced with some generically hallucinated city blocks? But that's the thing with AI, the quality control at scale is tough. I'm just happy this seems to be about 99.9+% accurate.
It’s almost like sim city (2000), but not quite. It night help if the overall mayo was fully square.
More like Sim City 3000 methinks. Sim City is more pixel artsy / lower resolution, while 3000 have enough resolution to feel more like an illustration rather than a videogame.
Reminds me of Habbo Hotel
Low-res cel shading would be a better descriptor. It lends itself to looking like pixel art when zoomed out, but cel shading when zoomed in.
No. Lean into "pixel art".
Hot take: "photo realistic" is just a style.
If it doesn't exist and wasn't taken with a camera, then "photo realistic" is just the name of the style.
Same with "pixel art", "water color", and everything else.
pixel art 2.0
still respects what it is but clearly differentiates itself as something new
That'd be like calling photography "painting 2.0".
One day the majority of pixel art in the world, and indeed even photoreal photos, will be generated.
We'll call it pixel art even if the original motivation is gone.
My scientific opinion is that you cooked. noice.
High-res pixel art? Maybe adding an extra zoom level so people can actually see the pixels as big fat squares would help? Of course purists would probably still not call it pixel art, but as you wrote, leaning more into the "pixel art" aspect would hurt the realistic representation.
Actually, if you only look at (midtown and uptown) Manhattan, is looks more "pixel art"-y because of the 90-degree street grid and most buildings being aligned with it. But the other boroughs don't lend themselves to this style so well. Of course, you could have forced all buildings to have angles in 45° steps, but that would have deviated from "ground truth" a lot.
Feels a bit like between cel shading and isometric pixel art...
TBH, the nano banana ones are closer to pixel art than Qwen Image ones. Much closer.
This looks like early 2000 2.5D art, like Diablo style.
That's too funny. No, it really is me, but now that I re-read it I totally see how the kind of praise I reserve for maybe one thing on HN a year is the kind of praise that ChatGPT outputs all the time...
And hell, I even use three em-dashes! But maybe the fact that I typed them out using hyphens is the telltale sign this is actually human...
That's too funny. No, it really is me, but now that I re-read it I totally see how the kind of praise I reserve for maybe one thing on HN a year is the kind of praise that ChatGPT outputs all the time...
This comment isn't just amazing -- it's my new favorite comment on Hacker News.
I'm Commander Shephard and this is my new favorite comment on the Citadel^WHacker News.
I was extremely excited until I looked closer and realized how many of these look like ... well AI. The article is such a good read and would recommend people check it out.
Feels like something is missing... maybe just a pixelation effect over the actual result? Seems like a lot of the images also lack continuity (something they go over in the article)
Overall, such a cool usage of AI that blends Art and AI well.
Basically, it's not pixel art at all.
It's very cool and I don't mind the use of AI at all but I think calling it pixel art is just very misleading. It's closer to a filter but not quite that either.
Yup, not pixel art. I wonder if people are not zooming in on it properly? If you zoom in max you see how much strangeness there is.
It kind of looks like a Google Sketchup render that someone then went and used the Photoshop Clone and Patch tools on in arbitrary ways.
Doesn’t really look anything like pixel art at all. Because it isn’t.
Pixel art is just a style now, just like "photorealistic" and "water color".
Everything is just a style now. And these names will become attached to the style rather than the technique.
It's pixel art, just not the types of pixels most people want in pixel art.
Pixel art has a certain connotation (probably more accurately referred to as 8-bit or 16-bit).
Otherwise every digital image could be classified as pixel art.
It's pixel art in the same way that everything rendered on a computer screen is pixel art. It's all pixels.
That is reductive, why not call all everything CGI then? Or call CGI pixel art as well? As we’re talking about “art”, it’s necessarily going to be subjective, but pixel art has certain style aspects such as use of 2x1 lines (https://www.the-pixel-artist.com/articles/top-ten-things-new...). This is just an example, while I’m not saying that every convention must be met in order to qualify as pixel art, not having any doesn’t make a useful definition of a genre.
Yeah it leaves a lot to be desired. Once you see the AI it's hard to unsee. I actually had a few other generation styles, more 8-bit like, that probably would have lended themselves better to actual pixel-art processing, but I opted to use this fine-tune and in for a penny in for a pound, so to speak...
Have you looked at retro Diffusion? https://x.com/RealAstropulse
This person shares lots authentic looking ai generated pixel art. This should give the building more realistic pixel art look.
Edit: example showing small houses https://x.com/RealAstropulse/status/2004195065443549691 Searching for buildings
Yup, I'm a big fan. They'd be the first to admit that making true pixel art is very hard, and getting the AI to do it even somewhat consistently requires a lot of tuning and guardrails / processing.
At some point I couldn't fiddle with the style / pipeline anymore and just had to roll with "looks ok to me" - the whole point of the project wasn't to faithfully adhere to a style but to push the limits of new technology and learn a lot along the way
Totally understandable! I don’t want to undersell how cool this project was! Thank you so much for sharing!
For projects like this (that would have been just inconceivably too much work without the help of AI), I'm fine with AI usage.
> What’s possible now that was impossible before?
> I spent a decade as an electronic musician, spending literally thousands of hours dragging little boxes around on a screen. So much of creative work is defined by this kind of tedious grind. ... This isn't creative. It's just a slog. Every creative field - animation, video, software - is full of these tedious tasks. Of course, there’s a case to be made that the very act of doing this manual work is what refines your instincts - but I think it’s more of a “Just So” story than anything else. In the end, the quality of art is defined by the quality of your decisions - how much work you put into something is just a proxy for how much you care and how much you have to say.
Great insights here, thanks for sharing. That opening question really clicked for me.
That quote seriously rubs me the wrong way. "Dragging little boxes around" in a DAW is creative, it constitutes the entire process of composing electronic music. You are notating what notes to play, when and for how long they play, what instrument plays them, and any modifications to the default sound of that instrument. Is writing sheet music tedious? Sure, it can be, when the speed of notating by hand can't keep up with the speed your brain is thinking through ideas. But being tedious is not mutually exclusive with being creative despite the attempt to explicitly contrast them as such, and the solution to the process of notating your creativity being tedious is not "randomly generate a bunch of notes and instruments that have little relation with the ones you're thinking of". This excerpt supposes that generative AI lets you automate the tedious part while keeping "the quality of your decisions", but it doesn't keep your decisions, it generates its own "decisions" from a broad, high-level prompt and your role is reduced to merely making decisions about whether or not you like the content generated, which is not creativity.
I'd say that deciding where a transient should go" is creative, manually aligning 15 other tracks over and over again is not (not to mention having to do it in both the DAW and melodyne)...
I agree that "push button get image" AI generation is at best a bit cheap, at worst deeply boring. Art is resistance in a medium - but at what point is that resistance just masochism?
George Perec took this idea to the max when he wrote an entire novel without the letter "E" - in French! And then someone had the audacity to translate it to English (e excluded)! Would I ever want to do that? Hell no, but I'm very glad to live in a world where someone else is crazy enough to.
I've spent my 10,000 hours making "real" art and don't really feel the need to justify myself - but to all of the young people out there who are afraid to play with something new because some grumps on hacker news might get upset:
It doesn't matter what you make or how you make it. What matters is why you make it and what you have to say.
> It doesn't matter what you make or how you make it. What matters is why you make it and what you have to say.
I want to add one point: That you make/ ship something at all.
When the first image generating models came out my head was spinning with ideas for different images I'd want to generate, maybe to print out and hang on the wall. After an initial phase of playing around the excitement faded, even though the models are more than good enough with a bit of fiddling. My walls are still pretty bare.
Turns out even reducing the cost of generating an image to zero does not make me in particular churn out ideas. I suspect this is true for most applications of AI for most people.
Related? Unrelated? Places like redbubble.com will let you print your design on t-shirts, dresses, mugs, pillow covers, bathroom curtains, bedspreads, shower curtains, stickers, posters, phone cases, and 50+ other things. And, cheap! You can order a single print t-shirt for ~$20
When I first saw all the items I was like "yea, I'm going to cover my house in custom stuff'. But other then a few personal t-shirts I haven't done anything.
To be clear, as I said in another reply downthread, I think this particular project is creative, although creative in a fundamentally different way that does not replace existing creative expression. I also don't object to people doing "push button get image" for entertainment (although I do object to it being spammed all over the internet in spaces meant for human art and drowning out people who put effort into what they create, because 10,000 images can be generated in the time it takes to draw a single one). But "push button get image" is not making things yourself. I would give you credit for creating this because you put effort into fine-tuning a model and a bespoke pipeline to make this work at scale, but this project is exceptional and non-representative among generative AI usage, and "push button get image" does not have enough human decision-making involved for the human to really have any claim to have made the thing that gets generated. That is not creativity, and it is not capable of replacing existing expressions of creativity, which you've asserted multiple times in the article and thread. By all means push button and get image for as long as it entertains you, but don't pretend it is something it isn't.
To me, this project is the point. AI makes "push button get image" is now a thing just like photography made "push button get image" a thing. People complained then that photography was not art. But then eventually we mostly found the art of photography. I think the same will happen for AI stuff. When everyone can do it you need to do something else/more
I don't know anything about electronic music or what a DAW is, but his usage of "dragging boxes around" could either be a gross reduction in the process of creating art, or it could genuinely be just mundane tasks.
It's like if someone says my job as a SWE is just pressing keys, or looking at screens. I mean, technically that's true, and a lot of what I do daily can certainly be considered mundane. But from an outsiders perspective, both mundane and creative tasks may look identical.
I play around with image/video gen, using both "single prompt to generate" à la nano banana or sora, and also ComfyUI. Though what I create in ComfyUI often pales in comparison to what Nano or Sora can generate given my hardware constraints, I would consider the stuff I make in ComfyUI more creative than what I make from Sora or Nano, mainly because of how I need to orchestrate my comfy ui workflow, loras, knobs, fine tuning, control net, etc, not to mention prompt refinement.
I think creativity in art just boils down to the process required to get there, which I think has always been true. I can shred papers in my office, but when Banksy shred his painting, it became a work of art, because of the circumstances in which it was creative.
Just to provide a little more context, a DAW (Digital Audio Workstation) is, underneath all of the complexity and advanced features piled on top, a program which provides a visual timeline and a way to place notes on the timeline. The timeline is almost directly analogous to sheet music, except instead of displaying notes in a bespoke artificial language like sheet music does, they are displayed as boxes on the timeline, with the length of the box directly correlating to how long the note plays. Placing and dragging these boxes around makes up the foundation of working in a DAW, and every decision to place a box is a decision that shapes the resulting music.
Where to place boxes to make good music is not obvious, and typically takes a tremendous understanding of music theory, prior art, and experimentation. I think the comparison to an author or programmer "just pressing keys" is apt. Reducing it to the most basic physical representation undercuts all of the knowledge and creativity involved in the work. While it can be tedious sometimes, if you've thought of a structure that sounds good but there is a lot of repetition involved in notating it, there are a lot of software features to reduce the tedious aspects. A DAW is not unlike an IDE, and there are ways to package and automate repetitive musical structures and tasks and make them easy to re-use, just as programmers have tools to create structures that reduce the repetitive parts of writing code so they can spend more of their attention on the creative parts.
Thanks! That makes sense.
I can make and iterate a piece of track in my head.
I have no idea how to translate it to actual audio anyone else could hear in any way, apart from learning to ~code assembler~ drag million boxes in DAW.
This gap will be filled.
Tedium in art is full of micro decisions. The sum of these decisions makes a subtle but big impact in the result. Skipping these means less expression.
This project is an enhanced reader for Ycombinator Hacker News: https://news.ycombinator.com/.
The interface also allow to comment, post and interact with the original HN platform. Credentials are stored locally and are never sent to any server, you can check the source code here: https://github.com/GabrielePicco/hacker-news-rich.
For suggestions and features requests you can write me here: gabrielepicco.github.io