Subpixel dithering !
1 bit per channel which mean that what you see is only 0 or 1 for each channel (R, G, B)
By using gaussian blur, the result is perceptually very good !
X compress the image a lot but this truly 1 subpixel ON / OFF
Not related to code... But when I use a LLM to perform a kind of copy/paste, I try to number the lines and ask it to generate a start_index and stop_index to perform the slice operation. Much less hallucinations and very cheap in token generation.
The current culture about AI & LLMs is that we are "only" memorizing the Web into model parameters and that a LLM is unable to "invent" new paradigms. Maybe we are under estimating what Unsupervised Learning & RL could provide. Re-inforcement learning is about exploring and finding new ways to accomplish a task and I see no limit here (except the computational resources needed).
Content semantic maps allows to visualize content of a website in a 2d space.
It allows new kind of usage : competitor analysis, conversion visualization, content gap analysis.
Subpixel dithering ! 1 bit per channel which mean that what you see is only 0 or 1 for each channel (R, G, B) By using gaussian blur, the result is perceptually very good ! X compress the image a lot but this truly 1 subpixel ON / OFF