[dllu]

I feel like getting LLMs to spit out an SVG is akin to getting a human artist to draw something by just reciting a list of coordinates. It's insanely hard and unnatural.

Image generation models nowadays can easily generate a photorealistic pelican riding a bicycle, where the bicycle has perfect structure. But it is, of course, only a raster image.

It seems that we're missing a kind of step to decompose an image into a list of instructions (say, SVG paths, or even brush strokes with a real brush) to reproduce it properly. Doing so would probably need a true understanding of the structure of the scene, which is something that AI still struggles with to this day.

[staticshock]

The pelican on a bicycle test is specifically about generating an SVG, fyi, not a raster.

[dllu]

I know. I'm just thinking about how to make AI create SVGs better... in theory, a sufficiently smart AI could "generate an image in its head", think about it, and then output the SVG paths to produce said image. Intuitively that would be somewhat closer to how human artists convert artistic visions into a sequence of arm movements while holding a brush (obviously, humans don't hold a fully formed, photorealistic image in the head while drawing, but rather vague concepts, but still).

[cherioo]

I don’t quite agree. Good human artist can visualize in their mind how to draw a picture, i think. Which i think is no different than LLM doing SVG drawing in their “head”. Anthropic’s recent post call this head-space “workspace”.

It just might feel foreign to human who does not have a SVG trained head-space.

[dllu]

Sorry, my comment was confusingly worded, but I am precisely thinking about the head-space thing you are talking about. My point is: the fact that image generation models can generate a perfect pelican riding on a topologically correct and highly detailed bike frame [1], suggests that AI does have the capability to correctly understand it internally. But something is disconnected when we just let an LLM directly output the SVG, and all the nuanced understanding is lost.

As such, modern LLMs still kind of suck at generating pelican bike SVGs with obvious errors:

* some omitted the bottom of the diamond which connects from the pedals to the rear wheel

* some added an extra connection from the pedals to the front wheel, making it impossible to steer

* none could align the head tube with the fork

* none added a correct offset to the fork

* none could generate the chain properly in a way that attaches to the two sprockets correctly

... whereas these errors do not appear in the raster image. (To be fair, the raster image has other weirdnesses, like the bird having arms and only one leg)

If we could harness the sort of internal thinking that must have happened when it generated the highly consistent raster image, but make it output SVG instead, we'd get much better SVGs. So this would be indeed "visualize in their mind how to draw a picture" before outputting the SVG.

[1] https://chatgpt.com/s/m_6a611c29c02481918fbb0f165eb83594

[0x000xca0xfe]

Image models that support text output like Image2, or general text models that can read images like Claude can vectorize raster images. But they aren't very good at it, doing it manually in Inkscape still produces better quality even when done by non-artists.