📱 thought for 2s.
Scratch image
2025-10
🧠 genius 🤍
Scratch image
2025-10
💭 Great read. And his workflow setup article is pretty good too: https://steipete.me/posts/just-talk-to-it
Scratch image
2025-10
💭 Using AI/LLMs to write code or a function is simple, when you attempt to build a project, then you have to worry about CI/CD, code reviews, quality and testing. Many “vibe coders” miss the point on it and are stuck in continuous debug mode. Streamline your dev flow with LLMs first! 2025-10
💭 PRD->ARCH->SCRUM->DEV->QA

Can we merge some different dimensions in the same perspective?

Scratch image
2025-10
⚙️ pm, sm, po, agent, dev, sm, qa, sm, pm, qa, sm, dev, qa, dev, qa. 2025-10
💻 Brilliant: https://www.alphaxiv.org/html/2510.01346 2025-10
💭 Something I’ve read online “Forget about prompts. It’s all about context. And yes, the more precise your context, the more relevant the model’s completion. And yes, it’s exhausting. Actually I believe gathering the most precise context is the core of your added value while working with LLMs” cannot agree more➕ 2025-10
😂 Hilarious & creative: “quicksort explained IKEA-style” https://idea-instructions.com/quick-sort/ 2025-09
✍️ as the only engineering person in the trade secrets, patents and trademarks class in law school, back in UT-Austin, I certainly do know and believe in the importance of the topic. If you ever interested, read the lawsuits around such topics. 2025-09
💭 BMAD method is a sign of where we are headed… 2025-09
💭 World doesn’t need AI generated movies for entertainment. 

World needs realistic computer games where user choices end up in different stories or timelines. It’s like a movie. That’s where AI-generation comes in. 

Gaming companies, I believe in y’all!

Always online on-time rendered. 2025-09

📱 rabbit r1 creations are pretty cool, someone made a Pokédex🤣love it 2025-09
📲 💭 hardware security “flaw” for Apple A17: https://github.com/JGoyd/Apple-Silicon-A17-Flaw 2025-09
🧐 I still find myself using Sonnet 4 more than GPT-5, for coding. 2025-08
🧠 Neural networks are becoming more efficient every day. The rate at which AI is advancing is both exciting and a bit scary. What happens when we can run GPT-5 level models on a smartphone? 2024-04