HNHacker News
TopNewBestAskShowJobs

jjcm

12,488 karma · joined August 13, 2010

Jacob Miller

Working on diffui.ai - diffusion for UI design.

Formerly Figma, Atlassian, and Microsoft.

AMA about diffusion, design systems, and webcomponents!

+1 808 366 1708 j@diffui.ai x.com/pwnies

submissionscomments
jjcm··on Mercury 2.5 LLM hits 770 tokens per second
I'm still sad that we haven't seen a new Taalas style chip a la https://chatjimmy.ai/. Smaller models are good enough now to make that insane burst of tokens so useful.
jjcm··on Claude Opus 5.5
Interesting. I just opened all 3 on a fresh install of firefox with no issues - can I ask what OS / do you have any js disabled / do you have webgl disabled?
jjcm··on No Sloptober
I just got back from a 2 week trip to China. I was in some of the more remote parts and my cell wasn't able to connect to their towers in that area, resulting in me not having the tourist VPN.

The side effect was I was fully cut off from my AI tools for those two weeks. I was coding "manually" during that time, and I think I accompished in two weeks what I previously had been able to do in a day. I'm not gonna lie, it was very, very stressful as a solo founder.

The industry moves so fast these days, that the only way to keep up with the speed is to leverage them. While I can appreciate the push of this to help your brain think independently/critically, the opportunity cost of a month of development without LLMs is too high a price to pay.

jjcm··on Claude Opus 5.5
I should give it a spin. 5.3 wasn't multimodal, but it looks like their flash release was. Thanks for the tip.
jjcm··on GPT-6 Sol and Luna
More image->html tests comparing Astra/Sol/Luna:

Design: https://image.non.io/78795662-8bfc-4e14-8d72-3738392aa6b3.we...

All 3 were given the same prompt to dynamically light these and to create the designs as a SPA with page transitions.

Astra: https://html.non.io/annui-astra

Sol: https://html.non.io/annui-sol

Luna: https://html.non.io/annui-luna

Luna gets the button wrong, and in the same way Grok/MiMo did. Looking into it more, it's because Luna actually searched my computer for similar builds, found the ones that I did for grok/mimo, and referenced their files. Astra is still the best by a significant margin in my eyes. Far more polish, better page transitions, effects that aren't overcooked and take into account the page. Better contrast.

jjcm··on Claude Opus 5.5
I've done GLM 5.3 previously here: https://news.ycombinator.com/item?id=49295420

Worth noting though that GLM 5.3 isn't multi-modal, so it doesn't have a vision layer. It is quite clever and hacks around it pretty effectively however. I'm running a deepseek 4 build now and will reply shortly with that.

jjcm··on Claude Opus 5.5
The designs are outputs from my own site. This has an overview of the process: https://diffui.ai/learn/new-site

The gist of it though is I take a prompt, expand it into a json blob specifying structure/palette/positioning of elements/etc, feed that into a diffusion model to output a few choices. Once I lock in a choice I take the pixel output + json blob and use it as input into followup pages. The json helps preserve the brand across multiple pages.

Once I have all the inputs I take their corresponding image+json blobs and feed them into an agent to create a web implementation.

For image models, diffui currently uses gpt-image-2.5, mai-image-2.6, and very, very rarely a post-trained version of flux 2 dev I've made for web design, though that one will be deprecated soon.

jjcm··on Claude Opus 5.5
Image->HTML tests:

Design: https://image.non.io/78795662-8bfc-4e14-8d72-3738392aa6b3.we...

Opus 5.5's output: https://html.non.io/annui-opus/

Overall it follows image designs quite well, but it did ignore asks to animate page transitions. Additionally it's the least performant of the ones I've built with Astra/Grok/MiMo, despite using a lot of the same code. I'd rate it just below Astra in capability, but still solidly second place.

For comparison with other drops this week + current #1:

Astra: https://html.non.io/annui/

MiMo: https://html.non.io/annui-mimo/

Grok 4.7: https://html.non.io/Annui-grok/

jjcm··on MiMo v2.6
That's fair - worth noting that none of them were instructed to make a mobile variant or to test the mobile size.
jjcm··on MiMo v2.6
Here's an image->html test for it using 2.6 Pro Ultraspeed, along with comparisons for grok 4.7 and Astra.

Design: https://image.non.io/78795662-8bfc-4e14-8d72-3738392aa6b3.we...

MiMo 2.6 Pro Ultraspeed (36min): https://html.non.io/annui-mimo/

Grok 4.7 (25min): https://html.non.io/Annui-grok/

Astra (19min): https://html.non.io/annui/

Overall this felt like the weakest of the three. Ultraspeed was fast as far as tokens per second goes, but it overthought quite a lot of things resulting it in having one of the longest build times. That overthinking didn't lead to better results either - note the statue with the cropped off arm. It's also the worst implementation of the dynamic lighting effect / displacement effect - the background especially has some significant distortion. Astra was the only one that seemed to understand that displacement should happen less the further something is in the distance.

Here's a vid of all 3 side by side with the source design: https://non.io/video/annui-comparison.mp4

jjcm··on Grok 4.7
Both implementations have that fwiw, it's just that the astra one is dialed back slightly.
jjcm··on Grok 4.7
For transparency, it took around 20M cache read + 1M input + 100k output tokens. API pricing for each:

Grok 4.7: $12.60

GPT Astra: $35.00

jjcm··on Grok 4.7
It's definitely gotten better at image->html workflows. Here's a test comparing Astra (currently SOTA at this) vs Grok 4.7:

Designs: https://image.non.io/78795662-8bfc-4e14-8d72-3738392aa6b3.we...

Astra's build: https://html.non.io/annui/

Grok's build: https://html.non.io/Annui-grok/

Additional prompt instructions: "Add scrolling clouds behind the statues. Dynamically light the statues based on mouse position. Use diffui to generate the normal maps/depth maps/roughness maps of the objects, and to separate out the assets on to different layers."

Overall I find these models are getting good at following image as a source of instructions, but their refinement of the output varies heavily between the models. Astra's final output feels more polished, has better visual contrast, and the animations between the pages are smoother. Grok also chose to light all of the background elements, which imo overcooks it a bit.

Still though, for the price it's a great starting point.

jjcm··on Qwen Image 2.1
It is, yes. This is for diffui.ai, which for now is closed source.
jjcm··on Qwen Image 2.1
My harness expands the prompt into a json representation that specifies layout much more rigorously, which is why you see such that amount of alignment between the two.

That internal json backing helps significantly when you want to maintain consistent design system components/patterns across multiple pages. The aligned layout is it working as intended.

jjcm··on Qwen Image 2.1
I run a prompt-to-ui design site that uses image models for the design process[1]. The text rendering especially makes this model deeply interesting to me, despite the license. Here are some tests using my harness comparing the outputs of gpt-image-2 and qwen 2.1:

https://html.non.io/qwen-comparison/

The text rendering definitely is much, much better than anything else on the open weights market right now. Small text fidelity is quite good. It seems like the text encoder however gets a little bit overloaded with larger prompts - note the presence of hex codes in the design output, those were inputs from the expanded prompt.

I'll be trying a post-training run on this for web design, it has some serious potential.

[1] diffui.ai

jjcm··on Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint
I'm seeing around 7.9GB of ram, 120 tokens/s on a 6000 pro blackwell.
jjcm··on Transitions.dev: UI transitions for AI agents
Regardless of the quality here, I see this as a failed business model. Subscribable libraries of skills / premade assets was effectively what Tailwind was doing, and they weren't able to make it work any longer.

I could see it maybe as a one-time purchase, but $90/yr/user is somthing I'd never grab.

jjcm··on Blizzard Revives 'StarCraft' Franchise with Open-World Sci-Fi Shooter
> a target release date of 2030

Don't get me wrong, I'm excited for a StarCraft shooter despite it not being StarCraft at all (I enjoy the universe). I wanted StarCraft: Ghost back in the day, and this feels like a taste of that dream.

That said, targeting a release of 2030 is wild to me. Look at the Astra game dev hype that's hitting twitter right now. Yes it's all filtered to the best possible examples / yes people are not one shotting these things, but still it's very clear that LLMs are beginning to be very capable at game dev in a way that doesn't look/feel like ass.

LLMs are going to become more capable over time. I think most of us can agree that eventually they'll be competitive with AAA developers given enough GPU cycles. I would argue that given rate of improvement, we're looking at that point coming in the next year or two.

To emphasize this point, a year ago GPT-5 got released. This was the type of game it would make: https://youtu.be/yTHo7tMborY?t=324

This is the type of game Astra is making: https://youtu.be/GuO_Eo34C8E?t=348

It's just starting to be capable of making assets in blender / unreal. Given this rate of advancement, what is the state going to be like 4 years from now? It really feels a bit like the "travelling to distant stars" problem, where at some point it's faster to wait for better engines than to leave now. I worry that by the time this gets released, the market is going to be flooded with fully custom AAA games that are hyperniche. 4 years out is a long, long time to wait, and it's a longer time to bet on success / market dynamics.

jjcm··on Ask HN: What are you working on? (September 2026)
https://makefaster.dev

I had a bunch of extra Fable credits, so I spent about $10k running autoresearch loops on 200 of the top github repos with frontends to try and speed up the frontend performance. I distilled them down into a leaderboard of the most common wins, and built out an autoresearch loop that takes those learnings and applies them to your repo.

Works with your existing claude/cursor/codex sub in a cute custom TUI.

I just submitted a Show HN for it: https://news.ycombinator.com/item?id=49687032

jjcm··on Tailwind Labs is joining Shopify
Really happy for Adam and the rest of the crew - they made a great product for a prior era, and hopefully this is a soft landing for them given their revenue stream dried up.

Tailwind at this point is a coding language and a brand with no direct product. It's a great acquisition when you want mind share and developer love, and are OK with there being no revenue involved. Shopify is a solid match for that.

jjcm··on ChatGPT Images 2.5
Indeed I am! Also nice catch - should have a fix up in the next hour. Hit me up with any other reqs - j@diffui.ai

Edit: confirmed that the download should be fixed now - should be properly a png now.

jjcm··on ChatGPT Images 2.5
It's both. I have several models loaded into diffui. Which one each prompt uses is determined based on user preference over time. Whichever image is currently at the top of any image node stack marks a win for the model that generated it, I assign each model an ELO score based on that, and I bias the chance each model is selected based on their ELO score. Right now gpt-image-2 is better than my own model, and it services around ~96% of the requests in diffui.

I'll also be adding in microsoft's mai-image-2.6 soon, but I need to update my SOC2 to add MS as a provider before I turn that on for other users. The full list of models in rotation is here: https://image.non.io/d53a9760-8b74-4386-b032-d59da2cd5319.we...

jjcm··on ChatGPT Images 2.5
Some UI tests with it:

Warcraft 3 style agentic dev interface: https://image.non.io/cd9ea5cd-8ed7-44e0-ad3f-480ff0e51875.we...

Overall it used the reference images I gave it a bit better than gpt-image-2. I noticed 2 had issues getting the blue button just right. 2.5 nailed it.

A "John Politics" meme site: https://image.non.io/d2922164-fa96-4d07-a141-2febadb02939.we...

Did very well modifying the pose while keeping the appearance of Glenn Powell. gpt-image-2 had a lot of the "fried" look for some of his skin in prior designs I did for johnpolitics.com

A cyberpunk inspired ramen website: https://image.non.io/8d5d8f10-0f0f-4d91-b5ea-33af7538b150.we...

Dark mode sites surfaced the fried look quite a bit in prior models, but this definitely looks better on that front. One thing that looks perhaps worse though is the microglyphs - note the teal lines to the bottom right of the ramen, they're kinda blurry / not straight.

Overall fixed some of the main issues / gripes I had with gpt-image-2

jjcm··on ChatGPT Images 2.5
I use gpt image 2 very heavily for my current project (ai UI design tool). The biggest improvement I'm seeing with this is in speed. I've generated around 50k images with gpt-image-2 via api, the the average latency has held at around 104s.

It's wild how much of a difference this is - images are coming in at around 35-40s. Very noticable, and makes a difference when you're iterating quickly: https://jjcm.org/gpt-image-2.5-speed.mp4

jjcm··on GPT-6 Astra on OpenRouter
Opus 5 is quite good, and has been the SOTA for this test (by my own subjective comparison) for the last few months. One thing it's always struggled with though is recreating smooth SVG curves/cutouts.

Another way to think about it is you could prompt Astra to put the building back in. You couldn't prompt Opus 5 to get the correct curvature of the line / cutout. That part has always been a huge struggle for models.

jjcm··on Artificial Analysis Intelligence Index v4.2
IMO one of the biggest losses of the OpenAI/Cursor breakup will be the loss of OAI models on CursorBench [1]. Their bench has always been one that most-fit my mental model of how good each of these models are. I find AA’s Intelligence index to often be out of alignment with my own subjective evals.

[1] https://cursor.com/evals

jjcm··on GPT-6 Astra on OpenRouter
It's ability to handle non-90 degree cutouts and shapes for web dev is one of the best I've seen. The vision model on this is VERY capable.

Here's an image design source of truth: https://image.non.io/78f4cd8b-2560-4643-9a51-96a89171f994.we...

And here's the page it build from it: https://image.non.io/e7d3a9e5-f9df-4fd8-b79f-1f90280f978f.we...

Note the flowing svg lines, and how accurately it recreated them. Here's Opus 5 for comparison - you can really see how while Astra really recreated the flow that was in the original design, opus only got the general vibe: https://image.non.io/dfe13de0-4487-431f-8b69-544ff3030dac.we...

One thing I will say is you are paying for quality. That site build cost $24 - extremely non-trivial for a simple frontend.

jjcm··on Salad Theory
Ontological discussions aside, I’ll share out my recipe for one of my favorite variations of a Caesar that embraces its origins. It’s one my friends usually request I make/bring.

It’s composed of four main parts, made in this order: the the meat, the croutons, the dressing, and the leaf.

For the meat we’re making a carne asada. You’ll want a flank or a skirt steak, or anything thinner with good marbling. For the marinade use olive oil as the base. I generally do around 2 cups of olive oil for a larger serving. Don’t sweat the exact amounts for a marinade. Add around 10% Apple cider vinegar, 10% red wine vinegar, and 10% lime juice. Use garlic, cumin, black pepper, salt, smoked chipotle peppers (you can also find them in an adobo sauce which works great - skip the sugar if you use the adobo), two spoonfuls of sugar.

Blend all of it together heavily, then pour it over the steak. I like a smoothie blender for this as they tend to work better for these small batches. IMPORTANT: don’t wash the blender, we’ll use the remains each time to influence the subsequent thing. Marinate the steak for at least 2 hours.

Next we’ll make the croutons. Store bought and homemade croutons are very different. For the marinade add a full clove of garlic into our used blender, black pepper, salt, two spoonfuls of nutritional yeast, a thumb sized portion of Parmesan, and rosemary. Add oil so it fills in the gaps in the blender. The goal is a paste, not a liquid. Blend it up. Cut up the bread into slices, then tear it into chunks. The tearing creates better texture. Put the paste in a bowl, smear it all around, dump the bread into, and toss it so it’s coating all the bread. Put it in the oven on a low broil (grill for those in Australia). Careful as it can burn easily. Check it every minute til browns, toss it, then put it back in. Repeat til the outer shell is dry but the inner part is still squishy. They shouldn’t be wet. Remove and chill for 10min before adding them.

For the dressing, half olive oil, 15% lime juice, 15% white wine vinegar, 10% pepitas, 10% Parmesan. This is your base. Add an egg yolk, black pepper, a squirt of mustard. Blend it up, taste it, adjust as needed. I skip the anchovy for the Mexican vibe this salad has as I don’t think it resonates well, but by all means go for it.

Grill the steak high heat.

For the leaf any salad leaf will do, but add radish slices as well.

Toss the leaf and the dressing, top with crouts, steak, pepitas, and Parmesan. Enjoy. Should take you 60-90min of prep.

jjcm··on CEO fired developers to make room for AI. Developers create open source AI CEO
I've actually had a lot of success with something similar for my startup. I'm a first time founder, and there are so many unknowns when you're getting started. I've been using a Hermes agent to act as my "boss", giving me three tasks to do each day / creating a memory of everything with the business.

It's really, really helpful. So many things I would have completely missed if it weren't for this. I don't think you need this synthetic org setup to get exec help at a small level. It really helped me with fundraising, incorporation, compliance, and just keeping me on track.

← PreviousPage 4 of 34Next →