Duraid πŸ”₯ (@duraidco) profile picture

I build stuff

London, England
Joined July 2011
Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)Photo by Duraid πŸ”₯ (@duraidco)
spend years trying to become a morning person 🚫 move to tokyo, go out during the day and work late nights βœ…οΈ
97
Duraid πŸ”₯ retweeted
me with gpt-5 in cursor rn: "build a $1B company. make no mistakes."
47
51
4
848
128,915
Weekend ship log: βœ… Added user auth & account management βœ… Upload & analyze internal user feedback (from reviews, support logs, survey responses, call transcripts, etc.) βœ… Shareable & downloadable user insights reports Insightful Cards started as a scrappy internal tool to automate my own user research. No auth. Now that people are signing up and asking for access, it was time to grow up a bit. Users can now create accounts and start generating insights on their own.
3
132
🚒 Just shipped: Insightful Cards A tool to turn noisy customer feedback into structured insight reports. PMs, UXRs, marketers β€” this one’s for you. Here’s what it does (and why I built it) 🧡
1
1
156
Built for: πŸ‘‰ Product teams prioritizing feature work πŸ‘‰ UXRs analyzing user sentiment πŸ‘‰ Marketers mining reviews for ad copy πŸ‘‰ Solo builders looking for fast customer insight
1
42
If you've ever dumped 500 comments into a spreadsheet and regretted your life choices… This might save you hours. Hope it helps β€” would love your feedback. 🧠 Try it here β†’ insightfulcards.com
34
UX Researchers: 1. What tasks are you streamlining with AI? 2. What tools are you using?
2
719
When you fix a bug with cursor
2
3
326
Every other tech product: CTRL + K to add link Meanwhile @figma:
1
109
Nvidia just dropped a 70B open source model that outperforms gpt-4o and sonnet 3.5 Link below πŸ‘‡
1
1
167
Giving it the old strawberry test (Nvidia vs. ChatGPT4o)
1
83
But can it code as well? Llama-3.1-Nemotron-70B-Instruct-HF scored 55% on aider's leaderboard, just behind plain llama-3.1-70b-instruct. For context, GPT 4o scored 72.9%.
74