jjcm

4 ms·

61.
by jjcm 2mo ago
That's a fair perspective of before/after. And yea, possibly this was just me curating my feed to a point where it's kind/pleasant. It truly is though - mine is free from politics and dinguses, it's just design and
62.
by jjcm 2mo ago
> The only way to prevent your tweets from being recommended to random people is to lock your account, and that makes it pointless to reply to people who don't follow you or to like their tweets because they don't see anything
63.
by jjcm 2mo ago
I actually really like that pricing strategy. It's very transparent
64.
by jjcm 2mo ago
There's a lot of negative about twitter here, especially related to timeline being shit. I'd like to give a different perspective - it's an algorithmic feed like anything else, it surfaces what you engage with. There are good
65.
by jjcm 2mo ago
Yea, the prompt was the expanded json the diffui harness uses. And re grabbing the designs, go for it. I expaneded out the designs here - feel free to copy them for your agent to build: https://diffui.ai/app/canvas&#x2F
66.
by jjcm 2mo ago
It's wild how much these agent interfaces have aligned UI-wise. There's almost zero difference between all of them.
67.
by jjcm 2mo ago
I was playing with this a bunch when it dropped a week ago. It's a big step up over their prior model, but it's nowhere near gpt-image-2 at least for high density UI design. Photos are a solved domain in my eyes, so the real quest
68.
by jjcm 2mo ago
Also tried Qwen Code. Tried it with a coding plan and it 401'd. Tried it with an api key I loaded up with pay-as-you-go credits and it 401'd. It did a self update and it worked for a single request (me just saying hi). Pasted in t
69.
by jjcm 2mo ago
Update: Tried using Qwen Desktop. It told me it succeeded, and linked me to the folder it created it in. That folder didn't exist, I asked Qwen where the files were, and it attempted to create them again. Stopped it, and asked what hap
70.
by jjcm 2mo ago
The visual web development / perceptionbench scores are very promising for image->html flows. Here are some test results. Original designs: https://image.non.io/257dc9cb-9e6b-4e00-8f12-23ea5e073649.we... These are f
71.
by jjcm 2mo ago
The quality is quite high, but an observation is the direction they're taking these models correlates heavily to the usage demand of China vs the West. Specifically, they're immensely focused on t2v for action / high effect s
72.
by jjcm 2mo ago
Oof, apologies. I have not done any work at all on mobile to make this work.
73.
by jjcm 2mo ago
You have my dota 2 aus guild beat by a couple of places: https://image.non.io/dad602c8-4774-47dc-b671-c515c67a0cf1.we... You definitely can do fun/creative things with LLMs, but they need a lot more steering/handh
74.
by jjcm 2mo ago
ty ty, though I do agree with azan that the imagery inside of it looks generated (as it is). That was really a test of whether or not newer models could implement box shapes that aren't provided by CSS. Opus 5 was one of the first that
75.
by jjcm 2mo ago
I think it's important to separate patterns from aesthetics. Patterns help us understand a new interface for sure, but aligned aethetics start to look mundane. That's why I was specifically calling out the color usage and typeface
76.
by jjcm 2mo ago
Yea. I have a very speed run version of how I made that one here: https://x.com/pwnies/status/2076755344471289898 But basically diffui's build step provides an API to create normal/roughness/depth m
77.
by jjcm 2mo ago
A few layers to it. When you prompt on my product I expand out that prompt into a semantic json blob that represents the style, theme, color palette, and subjects on the screen. I then feed that blob into the diffusion model itself. The rea
78.
by jjcm 2mo ago
Ha! Totally fair. 13 cents is 0% margin, which is what I'm targeting for individual plans. I literally just got my SOC2 compliance complete an hour ago, and I'm gonna roll out team plans/enterprise plans next week (unified bi
79.
by jjcm 2mo ago
> There are other aesthetics my brain associates with AI, like beige/cream colors, orange accents, and serif typefaces Something to consider regarding the narrow space AI-created designs align on: LLMs are trained to write consiste
80.
by jjcm 2mo ago
> every pen is deployable It will be interesting seeing how this evolves. Any free hosting solution tends to spiral towards abuse at some point. Still trying to solve it for my own stuff. That said, I do wonder how valuable something lik
81.
by jjcm 2mo ago
Yo that homepage zoom from the app screenshot straight into the experience was SMOOTH.
82.
by jjcm 2mo ago
> > All sufficiently capable models, open and closed, should go through mandatory safety testing. The problem with this is the cycles required to abliterate a model is significantly less than the cycles required to train a model. This
83.
by jjcm 2mo ago
This is my own tool, diffui. Thanks, though I will say I spent like... 7 minutes on this. Feel free to completely lift the design or the implementation.
84.
by jjcm 2mo ago
Shared here: https://news.ycombinator.com/item?id=49040344
85.
by jjcm 2mo ago
Not bad at all, and this is pretty consistent of what I've found from the current open source models. I haven't tried it with kimi 3 yet, that's on my todo.
86.
by jjcm 2mo ago
Right now generation is entirely at-cost (0% margin). Once I finish my SOC2 I'm going to release a enterprise/team license, which I'll charge a fixed 50% margin for. Long term the goal is for those enterprise licenses to be t
87.
by jjcm 2mo ago
Appreciate the kind words. Re copy for agent, I originally had a build step that was build into diffui itself, but what I found was that: 1. People wanted their own agent with its own skills/extensions to do the work 2. People wanted t
88.
by jjcm 2mo ago
IMO what makes things awesome is human hours invested. The ramen shop website above is pretty, but it's a veneer. It's not weird and awesome, it's just a representation of a site. I spent about... 7 minutes of my life making
89.
by jjcm 2mo ago
http://impeccable.style/ also just released their latest version which has some image->html conversion. They just released a couple days ago and haven't had a chance to try, but I suspect theirs is a bit more robust
90.
by jjcm 2mo ago
For which segment? The top one is just three vids I took of different builds I had, the lower ones which show the UI actually working are pure html/js/css. Some elements of them are actually interactive - ie you can switch the opt

More ›