Repo with 26 stars brought me the first serious lead today! It’s so powerful. Would you support me with more ⭐ 😎 Also - share your repo in a comment. Would love to support yours! https://t.co/QA23xwHxAu
Check out the prompt below and enjoy the output 😎
You are my personal brand strategist, creative director and senior frontend developer. Based on everything you know about me, create a personal homepage that feels like the website I would have if someone truly understood my work, ambitions, personality, projects and direction. This should not be a resume. This should not be a generic portfolio. This should feel like a sharp personal landing page for a founder, builder, expert or creator. Your job: 1. Infer my positioning. 2. Infer my tone of voice. 3. Infer what topics, projects and ideas I keep coming back to. 4. Turn that into a premium one-page website. Build it as a complete static website using HTML, CSS and JavaScript. The page should include: - a strong hero section with a memorable headline - a short personal positioning statement - a “What I’m really building” section - a “Projects, signals and proof” section - a “Recurring themes” section based on what I often talk about - a “What AI noticed about me” section - a “Now / Next” section - placeholder cards for projects, links and media - tasteful advanced animations - responsive design - dark mode - smooth scroll - microinteractions - polished visual hierarchy Style direction: - modern, premium, slightly futuristic - not startup-template generic - not corporate - not cheesy - more like a founder/operator personal site - strong typography, spacious layout, subtle motion Important: - Do not invent hard facts. - If you are unsure, label something as an assumption. - Make the copy specific, honest and useful. - Avoid AI clichés like “unlock your potential”, “revolutionize”, “game-changer”, “seamless”, “cutting-edge” unless truly justified. Generate the full code.
Small side experiment during my 90-day fitness project.
I asked a few AI models to generate how I might look after losing 9 kg.
Same photo → Grok, MiniMax, Gemini, ChatGPT.
Results ranged from
“basically the same guy”
to
“ready for a fitness competition”.
ChatGPT produced the most realistic one.
Comparison here: https://t.co/VKKmberD61
Day 2 building my “AI fitness operator” by #openclaw.
Tried logging lunch with a photo.
Sent a picture of spaghetti → system replied:
“rice with chicken and broccoli”.
😅
For a moment, I blamed the @MiniMax_AI model.
Turns out the model wasn’t the issue — the image never reached it.
Wrong tool in the pipeline + image input missing in the model config.
Two small fixes later, it reads nutrition labels correctly.
Food can now be logged by:
text, voice or photo.
Builder log: https://t.co/961TQcxRKa
Day 1 running my “AI fitness operator”.
A few early observations:
• morning analysis was too generic → updated prompt to include historical comparison
• agent forgot to ask for my weight → added explicit check-in
• food photo calorie estimates are pretty weak on the current model (Minimax 2.5)
One thing that worked nicely though:
I asked the agent to adjust my workout after a 30+ day break.
Through the Hevy API it reduced the weights ~15-20% and pushed the changes directly to the app.
Opened @HevyApp and the updated numbers were already there.
That felt like the system doing real work.
Wrote a short Day 1 log here:
https://t.co/eSmIspeYpl