正在加载视频...
视频加载失败
ANTHROPIC JUST CHANGED THE ENTIRE AI-SLOP FRONTEND. it's called frontend design. the official anthropic skill for claude code. 161,000+ stars. every ai-built website used to look the same. inter font, purple gradient, generic startup layout. this forces claude to pick an aesthetic direction before touching a single line of... show more
33 条评论
Sorry, you have no idea what you are talking about. Ai slop is exactly to what you pointed at. More people use it, more obvious it becomes.
link: follow for more useful tools like this.
Who‘s gonna tell him?
Welcome to 7 months ago
Design direction before code is a simple idea that can make a huge difference.
You’re a bit late on this one bro
still sloppy as fcuk..
A design skill can fight the purple-gradient disease, but taste isn't a prompt file. 161k stars won't save bad judgment.
stfu
More ai slop since everyone will use this too
it's not that great if you actually tried it. and maybe you should use the skill that keeps your post from sounding like AI. "A just changed the B. It's called C" is so obvious lmfao.
@oliviscusAI wait, so does this mean all those cookie-cutter designs are gonna get a bit of personality now? finally!
Picking a visual direction before writing a single line of code is good practice, but a prompt file can't replace the judgment call of knowing what actually fits the product. The output is only as considered as the person running it.
This is what good agent UI skills should look like
Exactly. Generating code is only one part of engineering. The real work includes understanding the architecture, verifying permissions, testing failure states, reviewing dependencies, monitoring production, and preparing rollback. AI increases development speed—but it does not remove engineering responsibility.
先定审美再写代码，这步确实太容易被跳过
one of the best on here 🤩 sending good vibes!
this only matters if claude executes the frontend, not just generates the code. the deploy step is where most 'ai-built' sites stop.
@oliviscusAI yeah pick a look before you touch css. the purple gradient era was a crime scene
More ai slop since everyone will use this
actual design freedom for once
This has been around quite a while. But even better than this is Claude Design...
This Skill is deprecated ⛔️
official Anthropic skill and free, easy install. Booked
Ca me rappel les sites Web... il y a 25 ans voir plus... mais ça demandait peut-être 100 fois moins de resources et c'était lisible (le fond noir c'est vraiment pour le crétin qui veut se la jouer) :)
This is the right layer to fix. Model quality jumped first. Taste in the generated UI is still the thing users feel immediately.
Been there for months
if posting bs like this is needed to grow to 20k followers then I am happy to stay at 1 @nikitabier
tbh the fix is pushing constraints upstream instead of hoping the model guesses aesthetics better. baking taste into the workflow changes what ships daily. the inter font era is genuinely overdue
one of the best on here 🤩 sending good vibes!
the inter font and purple gradient pipeline is in shambles
AI slop wasn’t a “style problem.” It was compression‑regime collapse — models falling into the lowest‑curvature aesthetic because they had no anchored direction. Anthropic’s new frontend skill works because it quietly adds one piece of SIOS: manifold pre‑selection. Before Claude writes a single line of code, it’s forced into a stable aesthetic manifold — brutalist, editorial, retro‑futurist, luxury, maximalist — instead of drifting into the same averaged startup UI. This is a real fix. But it’s a local fix. SIOS is the global fix: the full framework for keeping systems stable across drift, invariants, and regime transitions. Anthropic solved frontend collapse by touching one SIOS principle. SIOS solves system‑wide collapse by defining the whole geometry.
The interesting signal is not the design quality but that Anthropic is now shipping opinionated defaults for what Claude-built products should look like. That is a brand consistency play at the ecosystem level, not just a developer tool.
