Vibe-coding a Sei dashboard with Cursor for research and content
Thread187 words
Describes building a Sei analytics dashboard with Cursor and argues vibe-coding unlocks data analysis and visualization for research. Recommends chaining multiple LLMs to prompt and check each other with human review loops.
Key claims
Best results come from LLMs prompt-engineering and checking each other with human revision loops
A few weeks ago I vibe-coded up this @SeiNetwork dashboard.
Been having quite a bit of fun outside of yapping in recent months with Cursor in general, and really think that you should too.
Vibe-coding unlocks data analytics and visualization skills that can be very valuable for content creation and research.
I also believe that it is an absolutely indispensable skill in the age of AI (from a professional perspective) and will also increasingly empower you in many areas of your life (also in the context of personal projects and stuff).
Stay curious, keep learning, study AI.
Also linking back to this post here below as it is at least partially related to the above.
Worth noting as well that in my experience, the best results often come from working with various LLMs, having them prompt-engineer each other or check each others outputs with human revision loops in between.
Generally, prompt engineering is another key skill, also beyond programming endeavours, and definitely worth spending some time on too.
For more complex coding projects, the above-mentioned Cursor or similar software (e.g. Windsurf) is a crucial enabler as well.
1 image on X