Item — HN Reader

How are you doing this with Opus. Clearly I’m missing something. I always turn to ChatGPT when I need images because Opus typically refuses. I’ve tried Claude Code and Claude online in the past. I’m pretty sure neither created images for me and I thought this was because Anthropic was focused on code.

I guess I need to try harder. :)

3 comments

Images and build step were generated with my own tool (https://news.ycombinator.com/item?id=48995754 - it's why I'm often running these img->html tests).

Opus can't generate images since A\ doesn't have a diffusion model.

This is pretty interesting take on design and your trick to make the surface maps is pretty slick!

I did have one question about the tool, is it possible to set how many variations you want per step? I would rather be able to guide it manually at some steps where maybe I know pretty well what I want or just need minor tweaks, and then let it loose on others when really trying to experiment with an idea.

did you build your own diffusion model?

I have a few custom ones (a post-trained flux 2 checkpoint for web design and a image->metalness map generator that the build step can call for more advanced lighting situations), but gpt-image-2 is better than my own for design, so it's weighted much more heavily in outputs my tool generates. I think gpt-image-2 currently generates 99%+ of the design outputs on diffui

I believe they are testing giving it an image, which you can do in Claude code by dragging/dropping into the terminal or copy/pasting, and asking it to build the html equivalent.

You can just have claude code run the codex imagen via cli.