ANTHROPIC JUST PROVED MOST PEOPLE HAVE NO IDEA HOW TO PROMPT CLAUDE.
Their applied AI team dropped a 24 minute free workshop.
Not a creator who reverse engineered it.
Not a Reddit thread.
ANTHROPIC.
The people who wrote the weights.
And what they showed is uncomfortable.
There are 6 elements to a properly structured Claude prompt.
Most people are using 1.
Maybe 2.
That is not a skill issue.
That is an information issue.
And it has been quietly costing you every single day.
The outputs that felt slightly off.
The responses you had to rewrite 4 times.
The prompts that worked once and never again.
All of it traces back to the same 6 missing elements.
The people who watch this 24 minute workshop tonight will understand something about Claude that most daily users still do not know exists.
The people who skip it will keep getting 30% of what the tool is actually capable of and wonder why the results never quite land.
I watched it twice.
Then I built a Claude Skill that applies all 6 elements to every prompt automatically.
No more thinking about structure.
No more guessing what Claude needs.
The framework runs in the background every single time.
Full breakdown and skill setup is below.
Bookmark this now.
Watch the workshop first.
Then read the guide.
This is the one that compounds.
Follow @cyrilXBT for the exact prompt architecture, Claude skills, and systems I use to get outputs most people do not believe came from one person working alone.
INSTEAD OF WATCHING NETFLIX TONIGHT
Spend 30 minutes with this.
A speech by the Head of Anthropic's Coding Agents research team that teaches you more about vibe coding than 100 paid courses ever could.
This is not a tutorial.
This is the person who actually builds the system explaining how it works from the inside.
Most people learning to vibe code are learning the surface.
This goes underneath it.
30 minutes tonight.
The people who watch this will build differently tomorrow.
Completely free.
Bookmark this before you open Netflix 👇
APIs are cheap for building products. Not for vibe coding. Budget accordingly.
Tools that win are the ones that play well with your existing stack. Integration > isolation.
Design-specific AI flows are still early. But it's only a matter of time. Someone is gonna crack it soon.
Big takeaways: AI tools accelerate work but aren't magic. Human oversight still matters - especially for nuanced design. User friendliness beats raw power. The best tools don't make you read API docs. UX is now a competitive moat.
VS Code + MCPs: Most technical, most promising on paper, most disappointing in practice. Not ready for designers yet. What I could look into more for vibecoding but I still prefer Cursor by a mile.
Windsurf: Polished IDE with "bring your own tokens" - a game changer for cost control. Still needs better workflow integration. Another tool I haven't spent much time with. Also more oriented towards devs.
Figma Make: A no-brainer if you're already paying for Figma. Most designers are... employed or freelance. But it doesn't pull your Figma context naturally. I'm still struggling to use the design systems, in markdown files or using the "select library". Inconsistent output.
Framer Workshop: Brilliantly embedded in a tool designers already love. Loses focus over long sessions though. Wish there was a more complicated way to input markdown/long-term instructions.
Cursor: This is more vibecoding than vibedesign. I've tried it with Webflow/Figma MCP/Claude and a few other integrations. GitHub integration was a paradigm shift for a designer. Using OpenAI APIs directly finally made prompt engineering click.
Lovable: Best vibecoding/vibedesigning tool I tried. Fast results, generous quota of tokens, and it just got what I wanted. Best UX of the bunch. Cute name too.
Replit: Slooowww and clunky at first, but the companion app concept is revolutionary. The new design canvas is one to watch out for in the near future.
Vercel V0: Promising AI-generated UI, but lacked design finesse. It's improving, but the code generation ≠ design sense. We're using this at OHR, but the development team is handling most of the work while I pass on design in a traditional way.
N8N: Powerful for automation but not truly agentic. You still need some real coding chops to make it work. Not the "set it and forget it" I hoped for. It's improved, but I need to spend more time with it.
Put my 2025 favourite albums list on a substack. Gonna try and write some more about music I love to listen to, love to make, instruments, building, recording and playing effects and other techniques, all sorts of stuff. Follow along! #music#substack
https://t.co/tG4YS7LgXD